Preparation and Characterization of ZnTiO3/g‐C3N4 Heterojunction Composite Catalyst with Highly Enhanced Photocatalytic CO2 Reduction Performance

Herein, well‐crystallized ZnTiO3 particles are first prepared by hydrothermal method. A series of S‐scheme heterojunction photocatalysts of ZnTiO3/g‐C3N4 (referred to as ZTO/CN) with different mass ratios are synthesized by successfully doping ZnTiO3 in g‐C3N4 precursors and loading ZnTiO3 onto g‐C3N4 nanosheets by calcination. It is clearly found that the ZnTiO3 particles are successfully loaded on g‐C3N4 nanosheets by the X‐ray diffractometer, energy‐dispersive X‐ray spectra, and high‐resolution transmission electron microscopy images. Moreover, the specific surface area of 3.0% ZTO/CN is higher than that of pure g‐C3N4. Using triethanolamine as the hole sacrificial agent, the highest CO and H2 yields are achieved in the 3.0% ZTO/CN composite catalyst under the xenon lamp irradiation for 1 h. The generation rates of CO and H2 reach 15.19 and 5.77 μmol g−1 h−1, respectively, which are 2.9 and 4.1 times higher than that of pure g‐C3N4. The CO and H2 yields of the ZTO/CN composite catalyst show a trend of increasing and then decreasing with the increasing of ZnTiO3 content, which is due to the fact that excess ZnTiO3 can lead to a reduction of the effective heterojunction interface between ZnTiO3 and g‐C3N4, decreasing the transfer and separation efficiency of photogenerated electrons and holes and thus reducing the photocatalytic activity.


Introduction
With the rapid development of human society, traditional fossil fuels are being consumed rapidly.The traditional fossil fuels serve as nonrenewable energy sources and emit large amounts of carbon dioxide (CO 2 ) during combustion, exacerbating the energy crisis and global warming.[3][4][5][6][7] It is well known that solar energy is the most abundant renewable energy source on earth.However, its low energy density prevents us from using it directly.Photocatalysis, a technology that converts solar energy into chemical energy and has a role in environmental remediation, has attracted the interest of a wide range of researchers.Currently, the main photocatalysts are TiO 2 , [8,9] CdS, [10,11] graphene, [12,13] and graphitic carbon nitride (g-C 3 N 4 ). [14,15]However, these semiconductor photocatalysts have specific deficiencies in low solar energy utilization, low selectivity for target products, and high carrier recombination rates.Therefore, the development of photocatalytic systems with high activity and good selectivity remains a challenge.To solve this problem, researchers have done much research to improve the photocatalytic performance of semiconductor photocatalysts by loading precious metals, [16] doping other atoms, [17,18] vacancy modulation, [19] and constructing heterojunctions. [20,21]s a common two-dimensional semiconductor material, g-C 3 N 4 has been widely investigated in the fields of photocatalytic hydrogen production, pollutant degradation, and CO 2 Herein, well-crystallized ZnTiO 3 particles are first prepared by hydrothermal method.A series of S-scheme heterojunction photocatalysts of ZnTiO 3 / g-C 3 N 4 (referred to as ZTO/CN) with different mass ratios are synthesized by successfully doping ZnTiO 3 in g-C 3 N 4 precursors and loading ZnTiO 3 onto g-C 3 N 4 nanosheets by calcination.It is clearly found that the ZnTiO 3 particles are successfully loaded on g-C 3 N 4 nanosheets by the X-ray diffractometer, energy-dispersive X-ray spectra, and high-resolution transmission electron microscopy images.Moreover, the specific surface area of 3.0% ZTO/CN is higher than that of pure g-C 3 N 4 .Using triethanolamine as the hole sacrificial agent, the highest CO and H 2 yields are achieved in the 3.0% ZTO/CN composite catalyst under the xenon lamp irradiation for 1 h.The generation rates of CO and H 2 reach 15.19 and 5.77 μmol g À1 h À1 , respectively, which are 2.9 and 4.1 times higher than that of pure g-C 3 N 4 .The CO and H 2 yields of the ZTO/CN composite catalyst show a trend of increasing and then decreasing with the increasing of ZnTiO 3 content, which is due to the fact that excess ZnTiO 3 can lead to a reduction of the effective heterojunction interface between ZnTiO 3 and g-C 3 N 4 , decreasing the transfer and separation efficiency of photogenerated electrons and holes and thus reducing the photocatalytic activity.
[24] However, the low photocatalytic activity of g-C 3 N 4 is attributed to the high recombination rate of photogenerated carriers.To solve this problem of g-C 3 N 4 , researchers have proposed various strategies to improve its photocatalytic performance, such as morphology modulation, [25] atomic doping, [26] and heterostructure construction. [27]Among them, the construction of g-C 3 N 4 -based heterojunctions is an effective method.Because it is feasible and effective for the spatial separation of photogenerated electron-hole pairs, which can effectively accelerate the carrier transfer rate and significantly improve the light absorption capacity, the rational design of g-C 3 N 4 -based heterostructures can provide promising avenues for the creation of efficient visible-light-driven photocatalysts for environmental and energy applications.[33] However, the short wavelength absorption and high charge recombination rate of pure ZnTiO 3 photocatalysts are still problems of low photocatalytic performance.To address these drawbacks, attempts have been made to modify ZnTiO 3 with atomic doping, [34] but the modified ZnTiO 3 still cannot effectively utilize sunlight.Therefore, to effectively utilize sunlight and improve the photocatalytic performance, heterojunction structures are usually constructed with other materials such as ZnO [35] and rGO, [36] thus enhancing the photocatalytic performance of ZnTiO 3 .Currently, the ZnTiO 3 /g-C 3 N 4 heterojunction photocatalyst has been applied in pollutant degradation. [37]However, there are no studies about ZnTiO 3 /g-C 3 N 4 heterojunction photocatalysts for photocatalytic hydrogen precipitation and photocatalytic CO 2 reduction.Loading ZnTiO 3 onto g-C 3 N 4 to form a tight heterojunction interface can effectively reduce the photogenerated carrier recombination rate and thus enhance the photocatalytic activity.
In this study, a series of ZnTiO 3 /g-C 3 N 4 S-scheme heterojunction photocatalysts with different mass ratios were synthesized by loading ZnTiO 3 onto g-C 3 N 4 nanosheets by in situ growth method.And the structural and morphological characterization showed that ZnTiO 3 loaded on g-C 3 N 4 formed a tight heterojunction structure.The photocatalytic CO 2 reduction activity was tested under the xenon lamp irradiation for 1 h using triethanolamine as a hole sacrificial agent.The 3.0% ZTO/CN composite catalyst showed the highest CO and H 2 yields, with the generation rates reaching 15.19 and 5.77 μmol g À1 h À1 , respectively, which were 2.9 and 4.1 times higher than that of pure g-C 3 N 4 .The overall CO and H 2 production rates of the 3.0% ZTO/CN composite catalyst remained at 14.3 and 4.8 μmol g À1 h À1 after six cycles, which indicated the high stability of the prepared samples.The experimental results and a series of characterization data indicated that the composite of ZnTiO 3 and g-C 3 N 4 could promote the photogenerated electron-hole pair separation.The way of charge transfer under light illumination was also analyzed.This study provides new ideas for the design of g-C 3 N 4 -based composites for photocatalytic CO 2 reduction.and 0.4 g of NaOH and continue to stir for 1 h.The stirred solution was poured into a 50 mL PTFE reaction kettle and held at 200 °C for 24 h in an electric thermostatic blast dryer.After allowing the temperature to drop to room temperature, the hydrothermal solution was centrifuged and simultaneously washed three times alternately with deionized H 2 O and ethanol.The centrifuged samples were dried in a vacuum drying oven at 60 °C for 12 h.The dried products were poured into a crucible and heated to 700 °C in a muffle furnace at a heating rate of 7 °C min À1 for 2 h to obtain the desired ZnTiO 3 samples.

Preparation of ZnTiO 3 /g-C 3 N 4 Photocatalyst
2 g of melamine and 2 g of cyanuric acid are dissolved in 100 mL of deionized water at a constant temperature of 90 °C and stirred until they are completely dissolved.The melamine solution was added dropwise to the melanic acid solution with a rubber-tipped burette, and a white precipitate was formed during the dropwise addition.After the precipitate was allowed to come to room temperature, the resulting precipitate was centrifuged and washed three times with anhydrous ethanol and then dried overnight in an oven at 60 °C to obtain the g-C 3 N 4 precursor material.As shown in Figure 1, a certain amount of ZnTiO 3 powder was mixed with 2 g g-C 3 N 4 precursor in a mortar, and then the ground sample was transferred to a crucible, covered, and placed in a muffle furnace at a heating rate of 5 °C min À1 to 450 °C and held for 2 h.The ZnTiO 3 /g-C 3 N 4 sample was obtained when the muffle furnace was brought to room temperature.The ZnTiO 3 /g-C 3 N 4 composites in mass ratios of 0.02:1, 0.025:1, 0.03:1, and 0.035:1 were named 2.0% ZTO/CN, 2.5% ZTO/CN, 3.0% ZTO/CN, and 3.5% ZTO/CN, respectively.A pure g-C 3 N 4 sample was prepared in undoped ZnTiO 3 powder with the same conditions.

Characterization
The powder X-ray diffractometer (XRD) of the prepared samples was performed by X-ray diffractometer (XRD, UItima IV X) with Cu Kα radiation and at a scan rate of 10°s À1 .The morphology and microstructure of the materials were observed by scanning electron microscopy (SEM, Regulus 8220) and transmission electron microscopy (TEM, JEM 2100F).The specific surface area and pore size distribution of the samples were estimated by N 2 physisorption-desorption (BET, ASAP 2020 automated physisorption instrument, Mike's, USA) experiments.Fourier transform infrared (FT-IR) spectra of the samples in the range of 4000-400 cm À1 were obtained on a NICOLET 5700 infrared spectrometer.The UV-vis diffuse reflectance (DRS) absorption spectra of the samples were measured at 200-800 nm using a SHIMADZU UV-2600i spectrophotometer with BaSO 4 as a reference.The photoluminescence spectra (PL) of the samples were measured at an excitation wavelength of 350 nm from 300-800 nm using an F-4600 FL fluorescence spectrophotometer.Electron spin resonance (EPR) signals were analyzed on the EMXPlus at a microwave power of 20 mV.

Photocatalytic Reduction of CO 2
The photocatalytic activity of ZnTiO 3 /g-C 3 N 4 photocatalyst under visible light was tested.A 300 W xenon lamp (PLS-SXE 300D) from Beijing Perfect Light Company was used as the light source to simulate sunlight in this experiment.First, 30 mg of catalyst, 6 mL of mixed solvent (acetonitrile: water: triethanolamine = 3:2:1), and 5 mg of tris(2,2'-bipyridine) ruthenium (II) chloride hexahydrate were added into a quartz reaction tube of about 50 mL volume and sonicated for one minute to make the catalyst well dispersed in the solution.Second, the gas guide tube connected with the high-purity CO 2 (99.999%) gas cylinder was deepened below the liquid level of the reaction tube, and high-purity CO 2 gas was introduced for about 15 min to eliminate the air in the reaction tube to ensure the purity of CO 2 in the quartz reaction tube.Finally, the quartz reaction tube was sealed and illuminated with a xenon lamp for 1 h.After the light was finished, 1 mL of gas was extracted with a syringe in a WH500 gas chromatograph to analyze the gas in the reaction tube.The gas chromatograph was equipped with two detection channels, an FID detector and a TCD detector, to detect CO, CH 4 , H 2 , and other gases.

Photoelectrochemical Measurements
On the electrochemical analyzer (CHI660E), a platinum wire was used as the counter electrode, Ag/AgCl (saturated NaCl) was used as the reference electrode, and a fluorine-doped tin oxide conductive glass (FTO) coated with the sample as the working electrode.The photocurrent (IT) response, electrochemical impedance spectroscopy (EIS), linear scan (LSV), and Mott-Schottky (MS) of the catalysts were tested in a threeelectrode system with 0.5 M Na 2 SO 4 as the electrolyte solution.A 1.5 mL centrifuge tube containing 5 mg of the photocatalyst was added with 0.2 mL of ethanol and sonicated for 0.5 h to disperse the sample uniformly in the ethanol, then 10 μL of Nafion 117 perfluorinated resin was added to the solution and shaken well.Then, the prepared solution was applied on a 1 cm Â 1 cm FTO conductive glass sheet, and after the solution dried, nail polish was brushed around it.Finally, the FTO conductive glass was dried in an oven at 60 °C for 2 h.

XRD Analysis
The phase compositions and crystallinity of pure g-C 3 N 4 , pure ZnTiO 3 , and ZTO/CN composites catalyst were analyzed by X-ray diffractometry.The XRD spectra of pure g-C 3 N 4 , pure ZnTiO 3 , and ZTO/CN composites catalyst are shown in Figure 2a.It can be found that pure g-C 3 N 4 has two distinct diffraction peaks at 12.8°(100) and 27.7°(002), which are mainly due to the stacking of triazine units in the in-plane structure of g-C 3 N 4 and the stacking of interlayer aromatic chain segments. [38]The diffraction peaks at 2θ values of 24.0°, 32.8°, 35.4°, 40.6°, 49.0°, 53.9°, 61.9°, 63.3°correspond to (012), (104), ( 110), ( 113), ( 024), ( 116), (214), and (300) crystal planes of ZnTiO 3 (JCPDS NO. 26-1500), indicating that the ZnTiO 3 prepared by calcination has a good crystalline structure. [39]It is noteworthy that no other diffraction peaks were found in the XRD patterns of the ZTO/CN composites catalyst, which indicates that no other substances were doped during the composite process and the well-crystallized ZnTiO 3 has been successfully loaded on g-C 3 N 4 to form the ZTO/CN heterojunction composite photocatalyst.And because of the low ZnTiO 3 content, some of the diffraction peaks with a weak intensity of ZnTiO 3 are not detected in some composite samples.The characteristic diffraction peaks of g-C 3 N 4 are obvious in the ZTO/CN composites, which indicates that the ZTO/CN composite samples still maintain the basic structure of g-C 3 N 4 .To observe more clearly the diffraction peak changes of the heterojunction photocatalyst with the increase of ZnTiO 3 doping, a local amplification is performed.As shown in Figure 2b, with the increase of ZnTiO 3 content in the ZTO/CN composites, the intensity of two of the characteristic peaks, 32.8°and 35.4°is obviously increasing, while the intensity of the characteristic peak of g-C 3 N 4 at 27.7°is weakened to some extent.It is inferred that the presence of ZnTiO 3 limits the crystallization of g-C 3 N 4 to a certain extent, which further proves the successful construction of ZTO/CN composites.

FT-IR Analysis
To further investigate the molecular structure information of pure g-C 3 N 4 , pure ZnTiO 3 , and ZTO/CN composite samples, the prepared samples were characterized by FT-IR spectroscopy.As shown in Figure 3, for the FT-IR spectrum of pure g-C 3 N 4 , the absorption peak at 810 cm À1 is attributed to the bending vibrations of the triazine ring structure.The absorption peaks located at 1248, 1328, 1413, 1575, and 1634 cm À1 wave numbers are mainly caused by the stretching vibrations of the C─N and C═N bonds in the g-C 3 N 4 structure.A broad absorption peak exists at the position from 3100 to 3900 cm À1 , which is mainly caused by the stretching vibration of the N─H bond and the vibration of the O─H group of the water molecules adsorbed on the surface.For pure ZnTiO 3 , the absorption peaks at 438 and 530 cm À1 are caused by the typical stretching vibrations of the Ti─O and Zn─O groups in ZnTiO 3 . [40]The absorption peaks at 1632 and 3437 cm À1 are related to the bending vibration of the -OH group in the water physically adsorbed by the molecule. [41]The introduction of ZnTiO 3 on g-C 3 N 4 results in a slight change in the curve at 400 to 800 cm À1 compared to the spectrum of pure g-C 3 N 4 , which is attributed to the vibrational absorption of ZnTiO 3 .Notably, the FT-IR spectra of the ZTO/CN composites catalyst are similar to those of pure g-C 3 N 4 , indicating that the loading of ZnTiO 3 does not change the main structure of g-C 3 N 4 in the ZTO/CN composites catalyst.

SEM and TEM Analysis
The morphology of the sample was analyzed by scanning electron microscopy (SEM), as shown in Figure 4.A large sheet-like structure of g-C 3 N 4 can be seen in Figure 4a, which indicates the successful preparation of uniform g-C 3 N 4 nanosheets by calcination in a muffle furnace at 450 °C.In contrast, the ZnTiO 3 prepared at 700 °C shows irregular particles and strips (Figure 4b).4d-f.As shown in Figure 4d, the pure g-C 3 N 4 nanosheets undergo self-small focalization during calcination synthesis, which will lead to the reduction of their specific surface area and the limitation of their ability to trap light.Compared with the lamellar structure of pure g-C 3 N 4 , the 3.0% ZTO/CN composite catalyst shows the same overall lamellar structure as g-C 3 N 4 with some small ZnTiO 3 particles embedded in the g-C 3 N 4 lamellae, as shown in Figure 4e.This may be attributed to the fact that small particles of ZnTiO 3 are well dispersed on the g-C 3 N 4 nanosheets, effectively reducing the aggregation of g-C 3 N 4 during the selfassembled calcination synthesis.The results further indicate that the strong interaction between g-C 3 N 4 and ZnTiO 3 hinders the agglomeration of g-C 3 N 4 and ZnTiO 3 particles, respectively. [42]he HRTEM image in Figure 4f shows the successfully constructed heterostructure between ZnTiO 3 and g-C 3 N 4 .We can clearly observe the lattice stripe of ZnTiO 3 with a lattice spacing of 0.27 nm, which corresponds to the (104) crystal plane of ZnTiO 3 , indicating that the ZnTiO 3 prepared by the hydrothermal method has high crystallinity.
Furthermore, the results of energy-dispersive X-ray (EDX) elemental analysis of the selected regions marked in Figure 5 also confirm the successful loading of ZnTiO 3 on g-C 3 N 4 flakes. [43]herefore, the stronger interfacial interaction between g-C 3 N 4 and ZnTiO 3 facilitates the effective charge separation and thus improves the photocatalytic performance of ZTO/CN composites.

BET Analysis
To better understand the specific surface area as well as the pore size distribution of the samples, we used nitrogen physical  mesoporous structures in the samples.In addition, the maximum N 2 adsorption for all three samples is at a relative pressure close to 1.0 P/P 0 , which means they have more mesopores and macropores.The N 2 adsorption of 3.0% ZTO/CN composite catalyst has a certain degree of increase relative to g-C 3 N 4 .The BET specific surface areas of pure g-C 3 N 4 , pure ZnTiO 3 , and 3.0% ZTO/CN composite catalyst are 10.04, 1.82, and 18.73 m 2 g À1 , respectively.The specific surface area of the 3.0% ZTO/CN composite catalyst is almost 1.87 times higher than that of g-C 3 N 4 , which may be due to the loading of ZnTiO 3 that reduced the focusing of g-C 3 N 4 nanosheets.Moreover, the BJH pore size distribution data also clearly show that the 3.0% ZTO/CN composites have more abundant pores in the range of 10-80 nm than g-C 3 N 4 , which contributes to enhancing more reactive sites for reactant molecules and thus improve the photocatalytic performance.
The specific surface area, pore volume, and average pore size of the three samples are shown in Table 1.3.0% ZTO/CN has a specific surface area of 18.73 m 2 g À1 , while the pore volume is 0.48 cm 3 g À1 , and the average pore size is 41.78 nm.Compared with g-C 3 N 4 and ZnTiO 3 , the 3.0% ZTO/CN composite catalyst has a higher specific surface area. [44]It is well known that the higher specific surface area can provide more active sites, thus improving the catalytic performance of 3.0% ZTO/CN composite catalyst.

XPS Analysis
The composition and elemental valence of the compounds were analyzed by X-ray photoelectron spectroscopy (XPS).Figure 7 shows the XPS spectra of the 3.0% ZTO/CN composite catalyst.As shown in Figure 7, the measured full spectrum reveals the presence of C, N, and O elements, as well as a few amounts of Zn and Ti in the 3.0% ZTO/CN composite catalyst.In Figure 7b, the two peaks of 284.8 and 288.4 eV are shown in the C 1s spectrum, corresponding to the sp 2 C═C bond [45] and the sp 2 N─C═N bond in the nitrogen-containing aromatic structure.As shown in the N 1s spectrum in Figure 7c, we can see three peaks at 398.7, 400.5, and 404.4 eV.The peak at 398.7 eV corresponds to the sp 2 C─N═C bond in the triazine ring, the peak at 400.5 eV is attributed to the N─(C) 3 group, and the peak at 404.4 eV is related to the C─N─H group, [46] where the O 1s region shows a peak (Figure 7d) with a binding energy of 531.2 eV, which is attributed to the Ti─O bond and the Zn-O bond.The binding energies at 1020.9 and 1,044.1 eV in Figure 7e correspond to Zn 2p 3/2 and Zn 2p 1/2 , respectively.Figure 7f shows the Ti 2p spectrum with two peaks at 458.3 and 464.0 eV, which are the binding energies of Ti 2p 3/2 and Ti 2p 1/2 for Ti 4þ substances. [47,48]

UV-Vis DRS Analysis
The light absorption performance is an important observation of photocatalysts.To further explore the optical properties and  electronic energy band structure of g-C 3 N 4 , ZnTiO 3 , and ZTO/CN composites, we further investigate them using solid UV-vis diffuse reflectance.As shown in Figure 8a, the absorption wavelength of pure ZnTiO 3 nanoparticles is below 420 nm, and the g-C 3 N 4 nanosheets show an absorption edge at 470 nm.
The ZTO/CN composites catalyst does not see a significant enhancement in light absorption capacity in the 200-800 nm spectral range compared to pure g-C 3 N 4 , and this trend cannot be clearly seen with increasing ZnTiO 3 content.This is mainly due to the relatively low doping of ZnTiO 3 on the g-C 3 N 4 nanosheets, which makes g-C 3 N 4 not have much effect on light absorption.According to Tauc's law, [49] ðahvÞ The bandgap of the semiconductor is confirmed, where a, h, v, A, and E g are the absorption coefficient, Planck's constant, optical frequency, constant, and bandgap energy.The value of r is 2 due to the direct leap properties of g-C 3 N 4 and ZnTiO 3 by plotting the (ahv) 2 and hv plots.As shown in Figure 8b, the bandgap values of g-C 3 N 4 and ZnTiO 3 were obtained from the linear region intercepts in the plots, which were 2.86 and 3.04 eV, respectively, and the results are consistent with those reported in previous literature. [36,50]To determine the energy band variation of the catalysts, Mott-Schottky curves obtained from electrochemical tests were used to reveal the energy band positions of pure g-C 3 N 4 and ZnTiO 3 (as shown in Figure 8c,d).Both curves show a positive slope, so we consider pure g-C 3 N 4 and ZnTiO 3 as n-type semiconductors.In contrast to p-n heterojunctions, the charge carriers in such n-n type heterojunctions can be transferred between two different bandgap materials, thus allowing the charge carriers to be spatially separated without reducing the total number of charge carriers.From Figure 8c,d We can then calculate the g-C 3 N 4 and ZnTiO 3 valence band (VB) potentials as 1.83 and 2.79 V (vs NHE, pH = 7), respectively.

Photoelectrochemical Testing
To further validate the light absorption and conversion ability of the proposed ZnTiO 3 /g-C 3 N 4 heterojunction photocatalyst, transient photocurrent tests were performed.As shown in Figure 9a, the transient photocurrent images of pure g-C 3 N 4 , ZnTiO 3 , and ZTO/CN composites with alternating light without xenon lamp illumination and light for 20 s.It can be seen from Figure 9a that all samples show a significant photocurrent signal when light irradiation is performed, which diminishes immediately after the light is turned off.It indicates that all samples produce a photocurrent response during the intermittent on-off irradiation of the xenon lamp.In particular, the ranking of the photocurrent intensity under the same experimental conditions is 3.0% ZTO/CN > 2.5% ZTO/CN > 3.5% ZTO/ CN > 2.0% ZTO/CN > ZnTiO 3 > g-C 3 N 4 .The photocurrent intensity of the 3.0% ZTO/CN composite catalyst is much higher than that of the pure g-C 3 N 4 sample, indicating that the 3.0% ZTO/CN composite catalyst can achieve higher photogenerated electron-hole pair separation efficiency spatially, thus improving its photocatalytic performance for the photocatalytic reduction of CO 2 . [51]As shown in Figure 9b, we perform AC impedance test to further investigate the migration of photogenerated carriers.In general, the size of the radius of the Nyquist plot arc is related to the resistance of the electrode material and the material under test.Under the condition that only the test material is changed, the smaller the radius of the Nyquist plot arc, the lower the resistance of the test material and the higher the migration of photogenerated carriers in the material.The arc radius of the ZTO/CN composite sample is smaller than that of both pure g-C 3 N 4 and ZnTiO 3 , with the smallest arc radius for 3.0% ZTO/CN, indicating that the 3.0% ZTO/CN composite has a low electron transfer resistance and the fastest photogenerated carrier mobility.This result is consistent with the results of the transient photocurrent test.
It is well known that hydrogen production by electrolysis of water is a convenient and nonpolluting method, and the higher current density exhibited by the catalyst at the same potential indicates its higher hydrogen precipitation ability.We measured the electrocatalytic hydrogen precipitation performance of g-C 3 N 4 , ZnTiO 3 , and 3.0% ZTO/CN catalysts using LSV with 0.5 M Na 2 SO 4 as the electrolyte solution.As shown in Figure 9c, the 3.0% ZTO/CN catalyst exhibits superior hydrogen precipitation performance with higher current density at the same potential.The Tafel slope of 3.0% ZTO/CN (513 mV dec À1 ) is lower than that of pure g-C 3 N 4 (632 mV dec À1 ) and ZnTiO 3 (584 mV dec À1 ), as shown in Figure 9d, where the hydrogen precipitation kinetics can be indirectly responded by a linear fit of the Tafel curve.It is shown that 3.0% ZTO/CN provides more efficient interfacial charge transfer and enhances hydrogen precipitation performance.

PL Analysis
It is well known that the complexation of photogenerated electron-hole pairs can seriously affect photocatalytic reaction activity.To further investigate the efficiency of photogenerated carrier migration and recombination of the prepared samples, we performed photoluminescence tests at room temperature.Figure 10 shows the PL spectra of the pure g-C 3 N 4 , ZnTiO 3 , and ZTO/CN composites at an excitation wavelength of 350 nm.It can be seen from the PL spectra that the pure g-C 3 N 4 shows a distinct emission peak at 465 nm, which is attributed to the complexation of photogenerated electron-hole pairs in g-C 3 N 4 .The luminescence intensity of all ZTO/CN composites is increased compared with that of pure g-C 3 N 4 , which implies a higher recombination rate of photogenerated carriers due to the presence of a tight hetero structured interface between g-C 3 N 4 and ZnTiO 3 , which greatly facilitates the transfer of photogenerated charges from ZnTiO 3 to g-C 3 N 4 , resulting in a faster rate of interfacial complexation between the photogenerated electrons of CB in ZnTiO 3 and the photogenerated holes of VB in g-C 3 N 4 , promoting the separation of active charges between the CB of g-C 3 N 4 and the VB of ZnTiO 3 .54] These results are consistent with the photocatalytic CO 2 reduction activity of the ZTO/CN composites.

EPR Analysis
The electron paramagnetic resonance technique was used to capture unpaired electrons to identify vacancies.As shown in Figure 11, weak EPR signals associated with vacancies are detected at g = 2.004, indicating the successful formation of vacancies in the prepared samples.Compared with pure g-C 3 N 4 , 3.0% ZTO/CN appears to have a stronger signal, indicating a higher vacancy concentration.This is because the introduction of ZnTiO 3 leads to more nitrogen defects on the surface of g-C 3 N 4 , which induces the sample surface can provide more unsaturated sites.It is favorable for the formation of unpaired electrons and the photoinduced generation of charge carriers for a more efficient photocatalytic reaction.

Photocatalytic Activity Evaluation
The photocatalytic CO 2 experiments were carried out under the xenon lamp irradiation using triethanolamine as a hole sacrificial agent to investigate the photocatalytic performance of the prepared samples.As shown in Figure 12a, the CO and H 2 yields are 5.17 and 1.41 μmol g À1 h À1 for pure g-C 3 N 4 and 7.54 and 4.48 μmol g À1 h À1 for pure ZnTiO 3 , respectively.While the CO yield of ZTO/CN composites is increased compared with both pure g-C 3 N 4 and pure ZnTiO 3 .The 3.0% ZTO/CN composites catalyst has the highest CO and H 2 yields with the generation rates of 15.19 and 5.77 μmol g À1 h À1 , respectively, which are 2.9 and 4.1 times higher than those of pure g-C 3 N 4 .The CO and H 2 yields of ZTO/CN composites show a trend of increasing and then decreasing with increasing ZnTiO 3 content, which is due to the fact that excess ZnTiO 3 may lead to a reduction of the effective heterojunction interface between ZnTiO 3 and g-C 3 N 4 , decreasing the transfer and separation efficiency of photogenerated electrons and holes and thus reducing the photocatalytic activity.The above results indicate that the formation of heterojunctions is beneficial in improving the photocatalytic activity of g-C 3 N 4 .In addition to the determination of the photocatalytic CO 2 reduction activity, the reusability of the material is also very important in practical applications.Therefore, six cycles of photocatalytic CO 2 reduction experiments were performed for 3.0% ZTO/CN under the xenon lamp irradiation to examine the stability performance of the photocatalyst.Figure 12b shows that the CO and H 2 generation rates showed a slow decreasing trend in the six consecutive cycles of photocatalytic CO 2 reduction experiments, which may be due to a small loss of the samples when they were cleaned after each repetition so that the photocatalytic activity showed a decreasing trend.However, the overall   CO and H 2 production rates are maintained at 14.3 and 4.8 μmol g À1 h À1 , which indicates that the prepared 3.0% ZTO/CN composites have high stability.
To investigate the effect of light exposure time on the efficiency of photocatalytic CO 2 reduction of the samples, we performed light exposure and measured the CO and H 2 production rates for the 3.0% ZTO/CN photocatalysts at 1-4 h.From Figure 12c, we can see that the product CO and H 2 production rates do not increase exponentially with increasing time, which is attributed to the compounding of photogenerated electron-hole pairs.To distinguish the physical aspects synthesized with ZnTiO 3 and g-C 3 N 4 , we put 0.06 g ZnTiO 3 and 2 g g-C 3 N 4 into a research pot, mechanically mixed them, and tested their photocatalytic CO 2 reduction activity.As shown in Figure 12d, the CO and H 2 yields of 3.0% ZTO/CN heterojunction are significantly higher than those of the mechanical mixture of 3.0% ZnTiO 3 and pure g-C 3 N 4 , while the CO and H 2 yields of the mechanical mixture of 3.0% ZnTiO 3 and g-C 3 N 4 are close to those of pure ZnTiO 3 , indicating that the heterojunction structure plays a crucial role in the photocatalytic CO 2 reduction process role.
The synthesis process of photocatalysts can have an impact on the yield of CO 2 reduction products, which can lead to large errors in the results.Therefore, it is important to identify and eliminate carbon contaminants in the experiment.To ensure that the C in the product comes from CO 2 and not from carbon contaminants, we replaced the gas passed into the reaction tube with a stable argon gas.As shown in Figure 13, after 1 h of the shutdown reaction, the products of the photocatalytic CO 2 reduction of g-C 3 N 4 , ZnTiO 3 , and 3.0% ZTO/CN all showed a small amount of CO, and we guessed that this carbon contaminant could possibly come from the organic solvent in the reaction tube and ethanol from the product washing process, and the large amount of hydrogen produced is due to the reaction of the catalyst with H 2 O.

Photocatalytic CO 2 Reduction Mechanism
According to the experimental results of photocatalytic activity tests, the ZTO/CN composites exhibit more excellent photocatalytic activity compared to pure ZnTiO 3 and g-C 3 N 4 .This is mainly due to the successful formation of an S-scheme heterojunction interface between ZnTiO 3 nanoparticles and g-C 3 N 4 nanosheets.In UV-vis diffuse reflection and Mott-Schottky tests, we know that the valence band potential E VB = 1.83 V and conduction band potential E CB = À1.03V for g-C 3 N 4 and the valence band potential E VB = 2.79 V and conduction band potential E CB = À0.25 V for ZnTiO 3 .Combined with PL analysis results, we propose a possible charge transfer S-scheme mechanism, as shown in Figure 14.Under the illumination, electrons on VB of g-C 3 N 4 and ZnTiO 3 will be transferred from VB to CB due to the excitation of photons.The transfer of photogenerated charge from ZnTiO 3 to g-C 3 N 4 is greatly promoted by the coulomb force and built-in electric field, which makes the interfacial complexation rate between photogenerated electrons of CB in ZnTiO 3 and photogenerated holes of VB in g-C 3 N 4 faster and promotes the active charge separation between the CB of g-C 3 N 4 and VB of ZnTiO 3 .From the EPR test results, we know that with pure g-C 3 N 4 , 3.0% ZTO/CN appears to have a stronger signal, indicating a higher vacancy concentration.This is because the introduction of ZnTiO 3 leads to more nitrogen defects on the surface of g-C 3 N 4 , thus prompting the sample surface can provide more unsaturated sites, which is favorable for the formation of unpaired electrons and the photoinduced generation of charge carriers, resulting in more efficient photocatalytic reactions. [55,56]ll these results confirm the successful construction of the ZTO/CN S-scheme heterojunction structure and its important role in improving photocatalytic CO 2 reduction performance.

Conclusion
A series of ZnTiO 3 /g-C 3 N 4 S-scheme heterojunction photocatalysts with different mass ratios were successfully synthesized by doping ZnTiO 3 with g-C 3 N 4 by a hydrothermal method combined with calcination.The photocatalytic CO 2 experiments were carried out under the xenon lamp irradiation using triethanolamine as the hole sacrificial agent.The 3.0% ZTO/CN composites showed the highest CO and H 2 yields with the generation rates of 15.19 and 5.77 μmol g À1 h À1 , which were 2.9 and 4.1 times higher than those of pure g-C 3 N 4 , respectively.The CO and H 2 yields of the ZTO/CN composites showed a trend of increasing and then decreasing with increasing ZnTiO 3 content, which was due to the fact that excess ZnTiO 3 may lead to a decrease in the effective heterojunction interface between ZnTiO 3 and g-C 3 N 4 , which reduces the transfer and separation efficiency of photogenerated electrons and holes, thus decreasing the photocatalytic activity.The 3.0% ZTO/CN composite showed high stability in six cycles of photocatalytic CO 2 reduction experiments.The experimental results and a series of characterization data indicate that the composite of ZnTiO

Figure
Figure4cshows the SEM image of the 3.0% ZTO/CN composite catalyst, and we can see the ZnTiO 3 nanoparticles and nanorods attached to the g-C 3 N 4 nanosheets.These ZnTiO 3 nanoparticles and nanorods form a close heterojunction interface with the g-C 3 N 4 lamellae.TEM and HRTEM images were used to investigate the morphology of pure g-C 3 N 4 and 3.0% ZTO/CN composite catalysts, and the results are shown in Figure4d-f.As shown in Figure4d, the pure g-C 3 N 4 nanosheets undergo self-small focalization during calcination synthesis, which will lead to the reduction of their specific surface area and the limitation of their ability to trap light.Compared with the lamellar structure of pure g-C 3 N 4 , the 3.0% ZTO/CN composite catalyst shows the same overall lamellar structure as g-C 3 N 4 with some small ZnTiO 3 particles embedded in the g-C 3 N 4 lamellae, as shown in Figure4e.This may be attributed to the fact that small particles of ZnTiO 3 are well dispersed on the g-C 3 N 4 nanosheets, effectively reducing the aggregation of g-C 3 N 4 during the selfassembled calcination synthesis.The results further indicate that the strong interaction between g-C 3 N 4 and ZnTiO 3 hinders the
, we know that the flat-band potentials of g-C 3 N 4 and ZnTiO 3 are À1.13 and À0.35 V, respectively (vs Ag/AgCl, pH = 7).And the conduction band (CB) of the general n-type semiconductor is more negative than the flat potential by about 0.1 V. Thus, the CB potentials of pure g-C 3 N 4 and ZnTiO 3 are about À1.23 and À0.45 V, respectively, which are equivalent to À1.03 and À0.25 V compared to a normal hydrogen electrode (vs NHE, pH = 7).Previously we have known that the bandgap values of g-C 3 N 4 and ZnTiO 3 are 2.86 and 3.04 V from the formula

Figure 12 .
Figure 12. a) Yields of CO and H 2 for different photocatalysts; b) Recyclability of 3.0% ZTO/CN photocatalyst for photocatalytic CO 2 reduction under 300 W xenon lamp irradiation; c) Yields of CO and H 2 for 3.0% ZTO/CN photocatalyst under different illumination time; d) Mechanical mixing of ZnTiO 3 and g-C 3 N 4 for in situ synthesis of 3.0% ZTO/CN photocatalysts with yields of CO and H 2 .
3 and g-C 3 N 4 can improve photocatalytic activity by enhancing solar light utilization and promoting photogenerated electron-hole pair separation.This study provides new ideas for the design of g-C 3 N 4 -based composites for photocatalytic CO 2 reduction.

Figure 13 .
Figure 13.Yields of different photocatalysts under 1 h light in Ar (no CO 2 experiment).
The chemicals used in the experiments were of analytical grade and did not require further purification, and the deionized water was used throughout the laboratory process.Ethanol (CH 3 CH 2 OH), zinc acetate ((CH 3 COO) 2 Zn), sodium hydroxide (NaOH), and triethanolamine (C 6 H 15 NO 3 ) were purchased from Xilong Science Co., LTD.Tetrabutyl titanate (C 16 H 36 O 4 Ti) and melamine (C 3 H 6 N 6 ) were purchased from Shanghai Aladdin Biochemical Technology Co., LTD.Acetonitrile (C 2 H 3 N) and cyanic acid (C 3 H 3 N 3 O 3 ) were provided by Shanghai McLean Biochemical Technology Co., LTD. 2.1.Chemicals

Table 1 .
Specific surface area, pore volume, and average pore size of catalyst samples.