A p–n Junction by Coupling Amine-Enriched Brookite–TiO2 Nanorods with CuxS Nanoparticles for Improved Photocatalytic CO2 Reduction

Photocatalytic CO2 reduction is a promising technology for reaching the aim of “carbon peaking and carbon neutrality”, and it is crucial to design efficient photocatalysts with a rational surface and interface tailoring. Considering that amine modification on the surface of the photocatalyst could offer a favorable impact on the adsorption and activation of CO2, in this work, amine-modified brookite TiO2 nanorods (NH2-B-TiO2) coupled with CuxS (NH2-B-TiO2-CuxS) were effectively fabricated via a facile refluxing method. The formation of a p–n junction at the interface between the NH2-B-TiO2 and the CuxS could facilitate the separation and transfer of photogenerated carriers. Consequently, under light irradiation for 4 h, when the CuxS content is 16%, the maximum performance for conversion of CO2 to CH4 reaches at a rate of 3.34 μmol g−1 h−1 in the NH2-B-TiO2-CuxS composite, which is approximately 4 times greater than that of pure NH2-B-TiO2. It is hoped that this work could deliver an approach to construct an amine-enriched p–n junction for efficient CO2 photoreduction.


Introduction
The excessive emission of CO 2 resulting from the acceleration of industrialization and use of fossil fuel has caused global warming and serious environmental problems [1]. Photocatalytic CO 2 reduction, denoted as "artificial photosynthesis", is a promising technology for CO 2 conversion [2][3][4]. It is critical to develop efficient photocatalysts through surface and interface engineering [5,6]. Semiconductor TiO 2 has attracted extensive attention concerning CO 2 photoreduction owing to its stability, low cost, and low toxicity. However, TiO 2 is an n-type semiconductor with a wide band gap, and the disadvantages of limited light harvesting and poor electron-hole pair separation are not conducive to highly efficient photocatalytic CO 2 reduction [7][8][9].
Construction of a p-n junction is one promising approach to facilitate the separation of photogenerated carriers and improve the utilization of solar energy [10][11][12][13]. Fan et al. reported that 3D CuS@ZnIn 2 S 4 p-n heterojunctions with 2D/2D nanosheet subunits can promote the separation of photogenerated carriers and accelerate carrier transfer [14]. Yu et al. suggested that a CuO/TiO 2 p-n heterojunction can improve the separation efficiency of photogenerated electron-hole pairs [15]. It is well known that CuS and Cu 2 S are p-type semiconductors with narrow bandgaps [16,17]. When coupling Cu x S with TiO 2 , upon light irradiation, an internal electric field is established with the formation of a p-n heterojunction. Accordingly, the lower flat band potential of Cu x S allows for the transfer of photoexcited electrons from Cu x S to TiO 2 , while the holes diffuse from TiO 2 to Cu x S [18,19], which could suppress the charge recombination to achieve an efficient separation of electrons and holes during the photocatalytic process. In this regard, the formation of a TiO 2 /Cu x S p-n junction has potential for improving photocatalytic CO 2 reduction. However, most of the studies upon TiO 2 /CuS p-n junctions focus on the applications of photocatalytic pollution degradation and hydrogen (H 2 ) energy generation from water [19][20][21][22]. There are few works relevant to photocatalytic CO 2 reduction by designing TiO 2 /CuS composites. Recently, Lee et al. [23] reported a CuS x -TiO 2 film could effectively promote photogenerated charge separation for CO 2 photoreduction. It is still a challenge to rationally design CuS x -TiO 2 p-n junctions for efficient CO 2 photoreduction.
Due to the unique surface state and higher conduction band position of brookite TiO 2 compared to anatase and rutile [24,25], great potential has emerged in the field of photocatalytic CO 2 reduction. Liu et al. reported that defective brookite TiO 2 had the highest yield for CO and CH 4 production among the three TiO 2 polymorphs [26]. Subsequently, Peng's group studied exposed-crystal-face controlling [27], the construction of heterojunctions [28], and supported metal cocatalysts and dual cocatalysts [29,30] to improve the CO 2 photoreduction activity of brookite TiO 2 . In fact, the activation and adsorption of CO 2 are significant factors to enhance CO 2 photoreduction [31][32][33]. Surface amine modification has attracted great attention in this issue, because the amine groups can not only promote the adsorption and activation of CO 2 , but also coordinate with other metal ions to bind closely. Jin et al. [32] reported that surface amine modification enhances the activity of metal@TiO 2 photocatalysts. On the basis of the above backgrounds, in this work, amine-modified brookite TiO 2 nanorods coupled with Cu x S nanoparticles has been successfully fabricated. A significant p-n junction is formed between the NH 2 -B-TiO 2 -Cu x S interface, which effectively improves the transfer and separation of charge carriers. The composition and morphology of NH 2 -B-TiO 2 are characterized and the improved performance of photocatalytic CO 2 reduction is also discussed.

Materials Synthesis
Synthesis of amine-modified brookite TiO 2 (NH 2 -B-TiO 2 ): First, titanium glycolate precursor was synthesized based on our previous reports [34][35][36]; 5 mL tetrabutyl titanate (TBT) was placed into a round-bottomed flask containing 180 mL ethylene glycol (EG) and swirled with magnetic force for 1 h at 120 • C. The white powder material was washed five times with deionized water and once with anhydrous ethanol before being dried at 60 • C for 48 h. To make amine-modified brookite TiO 2 , 0.5 g of the as-prepared titanium glycolate precursor was disseminated in 35 mL of deionized water and 35 mL of ethylenediamine by ultrasonic treatment, and this mixture was uniformly transferred to a 100 mL Teflon-lined stainless-steel autoclave and then placed in an oven at 180 • C for 12 h. Following filtration and washing with deionized water 5 times and with anhydrous ethanol once, finally, the product was gathered after drying at 60 • C for 12 h.
Synthesis of brookite TiO 2 (B-TiO 2 ): Thus method is similar to the method used to prepare amine-modified brookite TiO 2 . Firstly, titanium glycolate precursor was synthesized and then 0.5 g of the as-prepared titanium glycolate precursor was dispersed in 64 mL of deionized water and 6 mL NaOH (1 mol/L) using ultrasonic treatment, which was evenly transferred to a 100 mL Teflon-lined stainless-steel autoclave and later heated at 180 • C for 12 h. Following filtration and washing with deionized water and anhydrous ethanol to pH = 7, the product was gathered and dried at 60 • C for 12 h.
Synthesis of Cu x S particles: First, 100 mg Cu(CH 3 COO) 2 ·H 2 O (0.5 mmol) was added to a round-bottomed flask containing 20 mL of anhydrous ethanol in an oil bath held at 80 • C under magnetic stirring, to which 42 mg of TAA was added, and kept at reflux for 4 h. The obtained product was centrifuged and washed numerous times with distilled water and anhydrous alcohol.
Synthesis of NH 2 -B-TiO 2 -Cu x S: As-synthesized NH 2 -B-TiO 2 (12.5, 25, 50 mg, respectively) and 10 mg Cu(CH 3 COO) 2 ·H 2 O were added to a round-bottomed flask containing 20 mL of anhydrous ethanol in an oil bath held at 80 • C under magnetic stirring, then we added 42 mg of TAA and kept at reflux for 4 h. Then, the suspension was washed with distilled water and distilled alcohol several times via centrifugation. Finally, the obtained product was dried at 60 • C for 12 h. Different ratios of NH 2 -B-TiO 2 -Cu x S composites were denoted as NH 2 -B-TiO 2 -Cu x S-n, where n represented the molar ratio of Cu x S, and the values of n were 8, 16, and 32, respectively.

Characterization of Photocatalysts
X-ray powder diffraction (XRD) measurements were carried out on an Ultima IV X-ray diffractometer with Cu Kα radiation in a range of 10-80 • and the scan rate was 10 • /min at 40 kV and 30 mA. Fourier-transform infrared (FT-IR) spectra were collected on a Nicolet iS10 IR spectrometer to analyze the chemical bonds and functional groups of the material. Scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) techniques on a JSM 2100 electron microscope operating at a 200 kV accelerating voltage were used to display the morphologies and elemental dispersion of the samples. X-ray photoelectron spectroscopy (XPS) spectra were measured using Thermo Scientific K-Alpha with an Al Kα X-ray sources (hν = 1486.6 eV, 12 kV 6 mA). All binding energies were calibrated through the C 1s peak at 284.8 eV. UV-vis diffuse reflectance spectrum (UV-vis DRS) was investigated on a Shimadzu UV-2550 spectrometer in a range of 200-800 nm. The photocurrent and electrochemical impedance measurements were performed on a CHI 760E electrochemical workstation, including a standard three-electrode system in 0.2 M Na 2 SO 4 solution, where Pt wire, Hg/HgCl 2 , and the as-prepared product were used as counter electrode, reference electrode, and working electrode, respectively. Mott-Schottky (M-S) and photocurrent decay plots were also carried out on a CHI 760E electrochemical workstation in 0.2 M Na 2 SO 4 solution. Among them, the preparation process of the working electrode involved weighing 10 mg of the sample in the sample tube and then adding 1 mL N, N dimethylformamide (DMF) and 20 µL Dupont Nafion membrane solution, which was then stirred for 30 min; after it was completely uniform, we used a pipette gun to transfer 30 µL of the mixed solution coated on the FTO conductive glass. Finally, it was dried in a vacuum oven.

Evaluation of Photocatalytic CO 2 Reduction
The photocatalytic reduction of CO 2 was evaluated under irradiation of 300 W xenon lamp and its wavelength was used to simulate sunlight in a gas-closed quartz reactor with a volume of 200 mL. Typically, 50 mg of the as-prepared catalyst was dispersed completely in 1 mL of deionized water in a glass Petri dish, which was then transferred to a quartz reactor with a bottom of 10 mL of deionized water. The reactor was bubbled with high-purity CO 2 gas for half an hour and the air inside was exhausted prior to illumination. Then, the reaction system was sealed and we turned on the xenon lamp. Subsequently, 1 mL of gaseous product was extracted from the glass reactor by a sampling needle every 1 h, and an irradiation duration of 4 h was applied and analyzed using a gas chromatograph (GC-7900, CEAULIGHT Beijing, China) equipped with an FID detector while N 2 gas served as the carrier gas. The reactor temperature was maintained at 25 • C and atmospheric pressure after starting the photocatalytic reaction.

Results and Characterization
The formation process of Cu x S particles supported on amine-modified brookite nanorods is shown in Figure 1. Firstly, the titanium glycolate precursor is synthesized by a simple reflux method. Subsequently, ethylenediamine is introduced to prepare aminemodified brookite TiO 2 nanorods via a hydrothermal method. Finally, Cu x S nanoparticles are deposited on the as-synthesized brookite TiO 2 nanorods by a refluxing method using copper acetate monohydrate and thioacetamide (TAA) as the precursors.   . It can be found that the peaks of brookite TiO 2 appear in all the as-prepared composites. However, the peaks of Cu x S can only be found in the NH 2 -B-TiO 2 -Cu x S 32% sample, which may be due to the strong diffraction peaks of brookite TiO 2 and low contents of Cu x S particles in this composite. The UV-Vis absorption (UV-DRS) spectra of the as-synthesized materials are displayed in Figure 2b. It became apparent that the pristine NH 2 -B-TiO 2 has an absorption edge at around 390 nm. However, the NH 2 -B-TiO 2 -Cu x S composites show red shift to the visible light region (400-800 nm), and the light absorption intensity gradually increases with the increase in Cu x S loading. Further, the Fourier-transform infrared (FT-IR) is used to analyze the chemical bonds and functional groups of the sample. Figure 2c shows FT-IR spectra of the NH 2 -B-TiO 2 and NH 2 -B-TiO 2 -Cu x S-16% samples, in which the brookite TiO 2 (B-TiO 2 ) (XRD in Figure 2d) is also prepared in the presence of NaOH for comparison. Indeed, the B-TiO 2 has no amine modification, while the NH 2 -B-TiO 2 and NH 2 -B-TiO 2 -Cu x S samples have two new peaks located at 1620 and~3400 cm −1 , respectively, which may be attributed to the N-H bending vibration and N-H stretching vibration, respectively [37,38].
To further confirm the composition of the sample, X-ray photoelectron spectroscopy (XPS) analysis was also performed to study the composition and surface chemical status of samples. Figure 3a shows the full XPS survey spectra of the NH 2 -B-TiO 2 , Cu x S, and NH 2 -B-TiO 2 -Cu x S samples. Among them, the full XPS survey spectra of the NH 2 -B-TiO 2 -Cu x S composite confirms the existence of N, Ti, O, Cu, and S elements. In Figure 3d, for N 1s in NH 2 -B-TiO 2 , the binding energy peak at 400 and 401.2 is attributed to the NH 2 group and N-H bonds [37,39], respectively, which suggests amine modification was achieved. After loading Cu x S, the N 1s peak shifted positively and it was proposed that the amines donate their lone pair of electrons on N atoms to Cu. For NH 2 -B-TiO 2 , the high-resolution XPS spectra of Ti 2p are shown in Figure 3b, and two tiny peaks observed at 458.6 and 464.3 eV are related to Ti 2p 3/2 and Ti 2p 1/2 , respectively, which indicates the presence of Ti 4+ [40,41]. Furthermore, Figure 3c exhibits two peaks located at 529.8 and 530.7 eV, which correspond to lattice oxygen and chemisorbed and dissociated oxygen, respectively [42][43][44].
For Cu x S, as shown in Figure 3e, the peaks at 932.6 and 952.7 eV belong to Cu 2p 1/2 and Cu 2p 3/2 of Cu + species, respectively [45,46], while the peak positions at 935.6 and 955.6 eV are consistent with Cu 2p 3/2 and Cu 2p 1/2 of Cu 2+ [47]. These characteristic peaks with a spin-orbit separation of 20 eV indicate the presence of Cu 2+ ions [16]. In addition, the other two satellite peaks of Cu + and Cu 2+ appear at 940.7 and 944.7 eV, respectively. Figure 3f displays the high-resolution spectrum at the S 2p region. The typical peak at 162.3 eV is from S 2p of S 2− [48,49]; moreover, a small peak at 169.1 eV is ascribed to SO 4 2− , which is due to S 2− being oxidized partially [50,51]. However, in the NH 2 -B-TiO 2 -Cu x S composite, all the binding energies of Ti, O, and N shift to higher regions, whereas Cu and S shift to lower binding energies compared with the pristine TiO 2 and Cu x S. It might be attributed to chemical interaction and electron transfer in the p-n junction formed at the interface of NH 2 -B-TiO 2 and Cu x S [52]. The morphology structures of the as-prepared materials are further determined by SEM. Figure 4a shows that the as-prepared NH 2 -B-TiO 2 is a rod-like structure with a diameter of 600-800 nm and a length of 2.5-3.5 µm. It can be observed that the microrods assemble with many tiny nanoparticles distributed on the surface. Figure 4b is an SEM image of Cu x S, in which the size and morphology are difficult to determine due to the phenomenon of particle aggregation. As shown in Figure 4c, Cu x S particles are dispersed on the surface of NH 2 -B-TiO 2 nanorods. In addition, energy-dispersive X-ray spectroscopy (EDX) mapping images of the NH 2 -B-TiO 2 -Cu x S composites show the presence and even distribution of Ti, O, N, Cu, and S elements on the surface of the nanorod in Figure 4d-i, which further confirms that NH 2 -B-TiO 2 and Cu x S are hybridized uniformly. In order to explore further CO 2 photoconversion efficiency, the photocatalytic CO 2 reduction performances of the as-obtained materials are evaluated in the presence of H 2 O using a gas chromatograph. As shown in Figure 5a, under the irradiation of a 300 W xenon lamp, only CH 4 was detected in the photocatalytic process of all samples, which may be because the prepared materials only meet the reduction potential of CO 2 reduction to CH 4 (E 0 = −0.24 eV). All the samples have the activity of photocatalytic reduction of CO 2 into CH 4 , except for the bare Cu x S. Among those samples, the brookite TiO 2 modified with amines (NH 2 -B-TiO 2 ) has superior capability to that of B-TiO 2 , indicating that amine modification has a positive effect on the reduction of CO 2 to CH 4 . With loading of Cu x S nanoparticles, the NH 2 -B-TiO 2 -Cu x S composites have enhanced performances for photocatalytic CO 2 reduction into CH 4 . In particular, the NH 2 -B-TiO 2 shows a low CH 4 production rate of about 0.73 µmol g −1 h −1 , while the optimized NH 2 -B-TiO 2 -Cu x S-16% sample has a yield rate of 3.34 µmol g −1 h −1 , which is 4-fold more than that of the pure NH 2 -B-TiO 2 . Those results suggest that the Cu x S could act as a cocatalyst in the amine-enriched B-TiO 2 -Cu x S composite for improvements in photocatalytic CO 2 reduction compared to pristine ones. Typically, to evaluate the stability of the as-synthesized photocatalyst, the photocatalytic CO 2 reduction activity with five-run cycling of the NH 2 -B-TiO 2 -Cu x S 16% sample is tested (Figure 5b). It can be found that the activity of CO 2 photoconversion to CH 4 is unstable, and the yield of CH 4 production gradually decreases, which could be attributed to the deactivation of the composite resulting from the oxidation of Cu x S.   Figure 6 shows the XPS spectra of Cu 2p of the NH 2 -B-TiO 2 -Cu x S 16% sample after the photocatalytic reaction. It can be observed that the valence state of copper displays obvious changes. The binding energies of 932.7 and 952.7 eV are ascribed to Cu 2p 3/2 and Cu 2p 1/2 , respectively, a typical peak location of Cu 2+ in CuS [14,20]. A weak satellite peak around 944 eV further indicates the presence of Cu 2+ [53]. These results indicate that Cu + was completely oxidized to Cu 2+ after the photocatalytic reaction in this photocatalysis system, which might weaken the cycling performance of the NH 2 -B-TiO 2 -Cu x S composite. As a matter of fact, as shown in the Mott-Schottky curve in Figure 7a-c, the slope of NH 2 -B-TiO 2 is positive, suggesting an n-type semiconductor [54]. At the same time, the slope of Cu x S is negative, suggesting that Cu x S is an p-type semiconductor. However, the NH 2 -B-TiO 2 -Cu x S-16% composite exhibits an inverted "V-shape", being a symbol of a typical p-n junction [55,56], which indicates that a p-n heterojunction could be constructed between the NH 2 -B-TiO 2 and Cu x S. In addition, the formation of the p-n heterojunction is further confirmed by valence band (VB)-XPS and core-level spectrum analyses. As shown in Figure 7d-f, the band alignment of NH 2 -B-TiO 2 and Cu x S between the NH 2 -B-TiO 2 -Cu x S heterojunction interface can be calculated according to the following: Equations (1)-(3) [57].
In the above equations, the ∆E VBO represents the valence band offset, which is the energy difference between the core energy level (E CL ) and the valence band maximum (E VBM ) in a pure material; meanwhile, ∆E Int CL illuminates the energy difference between the core levels. ∆E CBO represents the conduction band offset. Further, the band gap of the as-synthesized materials is calculated according to the Kubelka-Munk function in Figure 8a,b [58,59], and the band gap energies of pure NH 2 -B-TiO 2 and Cu x S are 3.14 and 2.04 eV, respectively. Based on the information reflected from the XPS and DRS analyses, Figure 8e reveals ∆E VBO = 2.38 eV and ∆E CBO = 1.28 eV for the NH 2 -B-TiO 2 -Cu x S nanocomposite. The VB-XPS spectra are verified, as shown in Figure 8c,d, as the valence band position for NH 2 -B-TiO 2 and Cu x S is 2.74 and 1.48 eV, respectively [60]. Therefore, the conduction band position of NH 2 -B-TiO 2 and Cu x S can be calculated as −0.4 and −0.56, respectively. In this regard, the formation of such a p-n junction and the resulting charge transfer are shown in Figure 8e. After contact, the Fermi levels of NH 2 -B-TiO 2 and Cu x S move down and up, respectively, until an equilibrium state is reached. When a built-in electric field between the NH 2 -B-TiO 2 and Cu x S interface is established, this allows for the electrons in Cu x S to migrate to NH 2 -B-TiO 2 while the holes in NH 2 -B-TiO 2 are transferred to the Cu x S. It is proposed that Cu x S as a cocatalyst promotes the efficient separation of photogenerated charge carriers. Generally, the separation efficiency of photogenerated electrons and holes has a significant impact on the photocatalytic performance [61][62][63]. Herein, photo/electrochemical measurements are carried out to study the charge transfer of the as-synthesized materials. In Figure 9a, the photocurrent density of the NH 2 -B-TiO 2 -Cu x S-16% composite is higher than pure NH 2 -B-TiO 2 , which indicates that the loading of Cu x S can effectively prevent recombination of the photogenerated electrons and holes. As shown in Figure 9b, the semicircle radius of the NH 2 -B-TiO 2 -Cu x S-16% sample is smaller than that of pure NH 2 -B-TiO 2 , implying that the charge carriers have a rapid transfer rate on the composite. Based on the above results and discussions, a possible CO 2 photoreduction process is proposed in Figure 9c. Before the photoreduction reaction, the surface amine modification is helpful for the adsorption and activation of CO 2 [31,37]. Under irradiation of a light source, n-type (NH 2 -B-TiO 2 ) and p-type semiconductors (Cu x S) generate photogenerated electrons in the conduction band (CB) and holes in the valence band (VB). The CB of NH 2 -B-TiO 2 is more positive than that of Cu x S, and the VB of Cu x S is more negative than that of NH 2 -B-TiO 2 . After contact between NH 2 -B-TiO 2 and the Cu x S interface, the built-in electric field is established, which promotes the migration of photoexcited electrons from the CB of Cu x S to NH 2 -B-TiO 2 and the migration of holes from the VB of NH 2 -B-TiO 2 to Cu x S, and these facilitate the separation and transfer of photogenerated electrons and holes. Further, the rate of reduction of CO 2 to CH 4 by the photoinduced electrons in CB of NH 2 -B-TiO 2 is improved. As shown in Table 1, the CO 2 photoreduction activity of the NH 2 -B-TiO 2 -Cu x S composite is higher than that of other TiO 2 -based binary and ternary composites reported previously.

Conclusions
In summary, an amine-enriched p-n junction upon the NH 2 -B-TiO 2 -Cu x S composite was successfully prepared. The modification of amine on the surface of the photocatalyst has a positive effect on the enhancement of CO 2 activity. The photocatalytic CO 2 reduction activity of amine-modified brookite TiO 2 is higher than that of amine-free modified brookite TiO 2 . Further, coupling different contents of Cu x S with NH 2 -B-TiO 2 , the NH 2 -B-TiO 2 -Cu x S-16% composite exhibits the greatest CH 4 yield rate of 3.34 µmol g −1 h −1 following 4 h of lighting, which is 4 times higher than that of pure NH 2 -B-TiO 2 . Combining the valence band XPS spectra with photo/electrochemical measurements, the formation of a p-n junction between the NH 2 -B-TiO 2 -Cu x S interface was confirmed. With the formation of such a heterojunction, the recombination of photogenerated electrons and holes is inhibited, thereby greatly improving the photocatalytic CO 2 reduction activity. It is hoped that this work could provide an approach to construct amine-enriched p-n junctions for efficient CO 2 photoreduction.

Conflicts of Interest:
The authors declare no conflict of interest.