Cu Nanoparticles Modified Step-Scheme Cu2O/WO3 Heterojunction Nanoflakes for Visible-Light-Driven Conversion of CO2 to CH4

In this study, Cu and Cu2O hybrid nanoparticles were synthesized onto the WO3 nanoflake film using a one-step electrodeposition method. The critical advance is the use of a heterojunction consisting of WO3 flakes and Cu2O as an innovative stack design, thereby achieving excellent performance for CO2 photoreduction with water vapor under visible light irradiation. Notably, with the modified Cu nanoparticles, the selectivity of CH4 increased from nearly 0% to 96.7%, while that of CO fell down from 94.5% to 0%. The yields of CH4, H2 and O2 reached 2.43, 0.32 and 3.45 mmol/gcat after 24 h of visible light irradiation, respectively. The boosted photocatalytic performance primarily originated from effective charge-transfer in the heterojunction and acceleration of electron-proton transfer in the presence of Cu nanoparticles. The S-scheme charge transfer mode was further proposed by the in situ-XPS measurement. In this regard, the heterojunction construction showed great significance in the design of efficient catalysts for CO2 photoreduction application.


Introduction
The rapid growth of atmospheric carbon dioxide (CO 2 ) concentration has attracted considerable attention due to its greenhouse effect on the global climate [1,2]. Developing artificial photosynthesis routes by reducing anthropogenic CO 2 emissions using solar energy instead of carbon-based fuels such as methane, methanol or carbon monoxide is still an intensive research topic in the environmental and energy field [2][3][4]. Since a TiO 2 photocatalyst was used to reduce CO 2 into methanol and formaldehyde by Inoue et al. in 1972 [5], most semiconductor photocatalysts, such as g-C 3 N 4 , CeO 2 , W 18 O 49 and Bi 2 WO 6 have received immense attention for CO 2 reduction under visible light irradiation, which occupies 44% of solar light [6][7][8][9][10][11]. Among the common photocatalysts, WO 3 material has gradually gained special scientific interest due to its relatively narrow bandgap (E g~2 .8 eV) and unique crystal structure containing a network of corner-shared octahedral units of [WO 6 ], which could theoretically utilize 12% of solar light and enhance charge migration in the catalyst [12]. Although WO 3 crystals exhibit various phases, few studies have focused on hexagonal phase WO 3 (h-WO 3 ) for photocatalytic CO 2 reduction [8,13,14]. h-WO 3 not only exhibits excellent visible light responsive properties according to its good photochromic ability [15], but also strong CO 2 adsorption at low-pressure due to the presence of ultramicro-sized tunnels [16]. It is thus considered as a potential visible-lightdriven (VLD) photocatalyst for CO 2 reduction. However, the application of photocatalytic WO 3 nanoflake film was synthesized by the solvothermal-calcination method. Before solvothermal growth, a thin seed layer was deposited onto the fluorine-doped tin oxide (FTO)-coated glass substrate (2.3 cm × 2.0 cm) by spin coating the precursor solution, which was made by dissolving 1.25 g of H 2 WO 4 in 30 wt% H 2 O 2 (8 mL), followed by annealing at 500 • C for 2 h in air. The H 2 WO 4 solution for solvothermal treatment was prepared by dissolving 1.25 g of H 2 WO 4 into 30 wt% H 2 O 2 (15 mL) and heating at 95 • C for 2 h. Nanoflake growth was achieved using 3 mL of H 2 WO 4 solution mixed with 0.5 mL of HCl (6 mol/L) and 12.5 mL of acetonitrile. A vertically oriented FTO-glass substrate was immersed into the above solution and placed within a Teflon lined stainless steel autoclave, which was then sealed and maintained at 180 • C for 3 h. The substrate was then rinsed with deionized water and dried in air. The WO 3 nanoflake film was obtained after annealing at 500 • C for 2 h in air.
The electrodeposition was performed in a standard three-electrode system. The FTO substrates (2.3 cm × 2.0 cm) were used as the working electrodes. The platinum sheet (1.0 cm × 1.0 cm × 0.2 cm) and Ag/AgCl electrode were used as the counter and reference electrodes, respectively. All of the chemicals were used without further purification. Before electrodeposition, the substrates were rinsed with distilled water and then cleaned by ultrasonic treatment in ethanol for 10 min. Cu 2 O was electrodeposited in 0.02 mol/L Cu(OAc) 2 aqueous solution (50 mL) containing 0.02 mol/L acetic acid (5 mL) using chronopotentiometry at −0.4 V with 0.5 s for 150 cycles. The Cu/Cu 2 O sample was obtained at −0.7 V under the similar conditions as above.
Cu 2 O/WO 3 and Cu/Cu 2 O/WO 3 films were synthesized by electrodeposition of Cu 2 O and Cu/Cu 2 O on the obtained WO 3 nanoflakes, respectively. Typically, the deposition process was conducted in Cu(OAc) 2 aqueous solution (0.02 mol/L, 50 mL) using chronopotentiometry electrodeposition with 0.5 s in 150 cycles, and the Cu 2 O/WO 3 and Cu/Cu 2 O/WO 3 films samples were obtained at −0.4 and −0.7 V, respectively. All the samples were then rinsed with deionized water and dried in air. Additionally, Cu 2 O and Cu/Cu 2 O films were prepared by the similar method, and Cu 2 O and Cu/Cu 2 O nanoparticles were broken off from the above samples by the ultrasound method. The dispersion solutions of Cu 2 O and Cu/Cu 2 O nanoparticles were then severally sprayed onto the WO 3 nanoflakes, and the obtained films were finally denoted as Cu 2 O/WO 3 m and Cu/Cu 2 O/WO 3 m samples.

Characterization
The crystal phases of the samples were recorded using an X−ray diffractometer (PANalytical X' pert PRO, Netherlands) with a Cu Kα irradiation source (λ = 0.154 nm) and 0.15 • /s scanning step. A scanning electron microscope (SEM, Nova NanoSEM 450, FEI) using the acceleration 300 kV voltage was used to characterize the morphology of the obtained products. Transmission electron microscopy (TEM) was obtained on a Tecnai G2 F20 S-TWIN electron microscope. Furthermore, high-resolution transmission electron microscopy (HRTEM) and Energy Dispersive X-ray spectroscopy (EDX) was employed and the corresponding fast Fourier transform (FFT) was evaluated by Gatan Digital Micro-graph software (Gatan Inc., Pleasanton, CA, USA). X-ray photoelectron spectroscopy (XPS) measurements were carried out at room temperature on a Thermo escalab 250Xi X-ray Photoelectron Spectrometer with a monochromatic Al Kα radiation (hv = 1486.6 eV). For XPS analysis, the samples without exposure to air were dried in N 2 flow gas and vacuum packed to avoid any impurity. All spectra were calibrated to the C 1s peak at 284.6 eV. The peak position was estimated using a fitting procedure based on the summation of Lorentzian and Gaussian functions using the XPSPEAK 4.1 program. UV-Vis diffuse reflectance spectra (DRS) were performed on a scan UV-Vis spectrometer (Cary 5000). The composition for the composites was determined by ICP-AES analysis using Thermo Scientic iCAP 6000 spectrometry. The photoelectrochemical test was recorded in a conventional three-electrode system by a CHI 660E electrochemical Nanomaterials 2022, 12, 2284 4 of 15 workstation (Chenhua, Shanghai, China). The photocurrents of the photocatalysts were measured at 0.0 V (vs. Ag/AgCl) in Na 2 SO 4 aqueous solution after being purged by N 2 under UV-visible light with an AM 1.5 G filter.

Photocatalytic Performance Tests
Photocatalytic CO 2 reduction activity with gaseous phase H 2 O was evaluated in a CEL-HPR100 stainless steel cylindrical vessel (Beijing China Education Au-Light Co., Ltd., China), and the light source was a PLS-SXE 300 Xenon arc lamp with a UV cutoff filter (λ > 400 nm). The compressed high purity CO 2 gas (99.995%) was passed across a deionized water bubbler, which generated a mixture of CO 2 and H 2 O vapor. After visible-light irradiation, the product yields and types in the gas phase were analyzed using a GC-7890II gas chromatograph (Beijing China Education Au-Light Co., Ltd., China), and the hydrocarbon product was further analyzed by 7890A-5975C GC-MS (Agilent Technologies Inc., Santa Clara, CA, USA). The experimental process is detailed in Supplementary Material S1.
The product selectivity (S) for CO 2 reduction was calculated with the following Equations (1)-(3): N CO , N CH4 and N H2 were on behalf of the yield of detected CO, CH 4 and H 2 molecules in the photocatalytic process of CO 2 reduction with H 2 O vapor.

Structure, Composition and Morphology
The XRD patterns of the as-prepared samples are presented in Figure 1 . This is because the percentage of exposed facets was estimated by the respective peak areas of the facets [41]. Hence, the WO 3 component in the bare and composite samples preferentially exposed (0 0 2) facets. Simultaneously, the diffraction peaks of Cu 2 O sample (marked with green dotted lines) matched perfectly with those of cubic phase Cu 2 O (PDF card No. 01-078-2076). To further explore the existence of Cu metal in the composite, the XRD pattern of Cu/Cu 2 O sample was studied, avoiding interference from the diffraction peak of hexagonal WO 3 , and the characteristic peaks at 43.4 and 50.6 • (marked with pink dotted lines) were attributed to metallic Cu (PDF card No. 03-065-9743). Moreover, Cu metal was proven to be present in the Cu/Cu 2 O/WO 3 composite. As a result, Cu and Cu 2 O were simultaneously electrodeposited onto hexagonal WO 3 , indicating the successful construction of the Cu/Cu 2 O/WO 3 composite.
X-ray photoelectron spectroscopy (XPS) measurements were carried out to elucidate the surface composition and chemical states of the elements. The survey XPS spectrum of the typical Cu/Cu 2 O/WO 3 sample indicated that the composite mainly consisted of W, Cu and O electrons. To gain further insight into the chemical bonding between W and other atoms in the composite, the high resolution XPS spectrum of W 4f (Figure 2a) was deconvoluted by Gaussian-Lorenzian analysis. The peaks at binding energies of 37.7 and 35.7 eV were ascribed to W (VI) state in tungsten oxide materials, while two distinct peaks at 34.5 and 36.4 eV were consistent with the values of W (V) oxidation state for all the samples [42,43]. The existence of W (V) was necessary to maintain the opening structure of hexagonal WO 3 [14]. However, the area ratios of W (V) and W (VI) in the composite were higher than those of the bare WO 3 sample (Table S1) due to the electroreduction process for Cu 2 O deposition. In the high-resolution Cu 2p XPS spectrum (Figure 2b), two conspicuous peaks were observed at binding energies of 952.3 eV for Cu (I) 2p 1/2 and 932.5 eV for Cu(I) 2p 3/2 [44]. Furthermore, the Cu and W diffraction peaks of the composite shifted slightly compared with those of pure samples, which was the reason that the intense interaction existed between WO 3 and Cu 2 O component in the composites, implying the formation of heterojunction. Cu LMM Auger spectra ( Figure 2c) were further performed to explore the chemical state of the Cu element, The Auger parameter can be calculated from the equation of α = E k (Auger electron) + E b (photoelectron). Here, E k is the kinetic energy, and E b is the binding energy. The Auger parameter values of the Cu (I) and Cu (0) in the Cu/Cu 2 O/WO 3 composite were determined to be 1848.7 and 1851.2 eV, respectively, which indicated the existence of Cu (I) and Cu (0) in the Cu/Cu 2 O/WO 3 sample [45,46]. X-ray photoelectron spectroscopy (XPS) measurements were carried out to elucidate the surface composition and chemical states of the elements. The survey XPS spectrum of the typical Cu/Cu2O/WO3 sample indicated that the composite mainly consisted of W, Cu and O electrons. To gain further insight into the chemical bonding between W and other atoms in the composite, the high resolution XPS spectrum of W 4f (Figure 2a) was deconvoluted by Gaussian-Lorenzian analysis. The peaks at binding energies of 37.7 and 35.7 eV were ascribed to W (VI) state in tungsten oxide materials, while two distinct peaks at 34.5 and 36.4 eV were consistent with the values of W (V) oxidation state for all the samples [42,43]. The existence of W (V) was necessary to maintain the opening structure of hexagonal WO3 [14]. However, the area ratios of W (V) and W (VI) in the composite were higher than those of the bare WO3 sample (Table S1) due to the electroreduction process for Cu2O deposition. In the high-resolution Cu 2p XPS spectrum (Figure 2b), two conspicuous peaks were observed at binding energies of 952.3 eV for Cu (I) 2p1/2 and 932.5 eV for Cu(I) 2p3/2 [44]. Furthermore, the Cu and W diffraction peaks of the composite shifted slightly compared with those of pure samples, which was the reason that the intense interaction existed between WO3 and Cu2O component in the composites, implying the formation of heterojunction. Cu LMM Auger spectra ( Figure 2c) were further performed to explore the chemical state of the Cu element, The Auger parameter can be calculated from the equation of α′ = Ek (Auger electron) + Eb (photoelectron). Here, Ek is the kinetic energy, and Eb is the binding energy. The Auger parameter values of the Cu (I) and Cu (0) in the Cu/Cu2O/WO3 composite were determined to be 1848.7 and 1851.2 eV, respectively, which indicated the existence of Cu (I) and Cu (0) in the Cu/Cu2O/WO3 sample [45,46]. The morphologies of the as-obtained WO3, Cu2O/WO3 and Cu/Cu2O/WO3 powders were visualized using SEM images. As shown in Figure 3, all the samples exhibited the typical sheet morphology with various sizes. Furthermore, the SEM image in Figure 3a  The morphologies of the as-obtained WO 3 , Cu 2 O/WO 3 and Cu/Cu 2 O/WO 3 powders were visualized using SEM images. As shown in Figure 3, all the samples exhibited the typical sheet morphology with various sizes. Furthermore, the SEM image in Figure 3a showed that the nanoflake surface was rough and porous. As shown in Figure 3b,c, both of the composites kept uniform sheet morphology, and the sheet thicknesses after electrodeposition of Cu 2 O and Cu/Cu 2 O had no obvious changes compared with that of bare WO 3 . Although there was no obvious difference in Cu content between Cu 2 O/WO 3 and Cu/Cu 2 O/WO 3 according to the ICP-AES analysis (Table S2), the nanoflake surface of Cu/Cu 2 O/WO 3 exhibited distinct embossment (Figure 3c). The morphologies of the as-obtained WO3, Cu2O/WO3 and Cu/Cu2O/WO3 powders were visualized using SEM images. As shown in Figure 3, all the samples exhibited the typical sheet morphology with various sizes. Furthermore, the SEM image in Figure 3a showed that the nanoflake surface was rough and porous. As shown in Figure 3b,c, both of the composites kept uniform sheet morphology, and the sheet thicknesses after electrodeposition of Cu2O and Cu/Cu2O had no obvious changes compared with that of bare WO3. Although there was no obvious difference in Cu content between Cu2O/WO3 and Cu/Cu2O/WO3 according to the ICP-AES analysis (Table S2), the nanoflake surface of Cu/Cu2O/WO3 exhibited distinct embossment (Figure 3c).  Micro-structures of the obtained samples were investigated by TEM observation. Figure 4 revealed that the typical nanoflake-like morphology was observed with a sharp edge and clean boundary. In Figure 4a, the WO 3 nanoflake had a porous surface morphology. The fringe spacings of 0.634 and 0.366 nm were discovered (inset of Figure 4a), which were consistent with the (1 0 0) and (0 0 2) lattice planes of hexagonal WO 3 , respectively. In Figure 4b, Cu 2 O/WO 3 still kept a nanoflake-like morphology with smaller porous sizes compared with that of WO 3 . The lattice fringes with spacings of 0.246 and 0.213 nm were distinctly found (inset of Figure 4b Figure S1) obeyed a logical normal distribution with an average diameter of 5.6 ± 1.1 nm ( Figure S2). According to the elemental mapping images in Figure 4d, W, Cu and O elements were uniformly distributed in the Cu/Cu 2 O/WO 3 sample, and the Cu element appeared on the surface, as well as around the nanoflakes. The above results demonstrate that Cu 2 O was uniformly electrodeposited on the surface of the WO 3 nanoflake, and metallic Cu also existed in the Cu/Cu 2 O/WO 3 sample.
The light absorption property is crucial for the utilization efficiency of solar energy for the catalyst. According to the UV-Vis diffuse reflectance spectra ( Figure 5), all the composites showed strong absorption in the visible light region, indicating the feasibility of utilizing visible light for CO 2 photoreduction. The band gaps (E g ) of the samples were estimated by the plots of (αhν) 2 versus photo energy (hν) ( Figure S3 and in the selected area (framed in red, corresponding enlarged view in Figure S1) obeyed a logical normal distribution with an average diameter of 5.6 ± 1.1 nm ( Figure S2). According to the elemental mapping images in Figure 4d, W, Cu and O elements were uniformly distributed in the Cu/Cu2O/WO3 sample, and the Cu element appeared on the surface, as well as around the nanoflakes. The above results demonstrate that Cu2O was uniformly electrodeposited on the surface of the WO3 nanoflake, and metallic Cu also existed in the Cu/Cu2O/WO3 sample.   The light absorption property is crucial for the utilization efficiency of solar energy for the catalyst. According to the UV-Vis diffuse reflectance spectra ( Figure 5), all the composites showed strong absorption in the visible light region, indicating the feasibility of utilizing visible light for CO2 photoreduction. The band gaps (Eg) of the samples were estimated by the plots of (αhν) 2 versus photo energy (hν) ( Figure S3 and Table S2), and the calculated Eg values of WO3 and Cu2O were 2.86 and 2.05 eV, respectively.

Photocatalytic Performance of CO2 Reduction
The photocatalytic CO2 reduction experiments were carried out under vis irradiation (> 400 nm) for the as-prepared catalysts. The photocatalytic perform CO2 reduction with water vapor in Figure 6a showed that no products were g over the WO3, Cu2O and Cu/Cu2O samples. Conversely, CO, CH4, O2 and H2 were obtained for the Cu/Cu2O/WO3 and Cu2O/WO3 composites, and no other such as HCHO or CH3OH were detected by either GC or GC-MS analyses. Furth the product yields decreased for the mechanically dispersed Cu2O/WO3 m sam pared with those of the Cu2O/WO3, indicating that the loosely contacted inter insufficient for photocatalytic CO2 reduction. Additionally, the product selectivity reduction was analyzed for the Cu/Cu2O/WO3 and Cu2O/WO3 catalyst. Typically and SCO values of Cu2O/WO3 were calculated to be about 0.1% and 94.5%, which strated that CO2 reactant was thermodynamically favorable to form CO by the t tron reduction pathway. However, the ScH4 and SCO values of Cu/Cu2O/WO3 we

Photocatalytic Performance of CO 2 Reduction
The photocatalytic CO 2 reduction experiments were carried out under visible-light irradiation (>400 nm) for the as-prepared catalysts. The photocatalytic performances of CO 2 reduction with water vapor in Figure 6a showed that no products were generated over the WO 3 , Cu 2 O and Cu/Cu 2 O samples. Conversely, CO, CH 4 , O 2 and H 2 products were obtained for the Cu/Cu 2 O/WO 3 and Cu 2 O/WO 3 composites, and no other products such as HCHO or CH 3 OH were detected by either GC or GC-MS analyses. Furthermore, the product yields decreased for the mechanically dispersed Cu 2 O/WO 3 m sample compared with those of the Cu 2 O/WO 3 , indicating that the loosely contacted interface was insufficient for photocatalytic CO 2 reduction. Additionally, the product selectivity for CO 2 reduction was analyzed for the Cu/Cu 2 O/WO 3 and Cu 2 O/WO 3 catalyst. Typically, the S CH4 and S CO values of Cu 2 O/WO 3 were calculated to be about 0.1% and 94.5%, which demonstrated that CO 2 reactant was thermodynamically favorable to form CO by the two-electron reduction pathway. However, the Sc H4 and S CO values of Cu/Cu 2 O/WO 3 were calculated to be 96.7% and 0.0%, indicating that the metallic Cu promoted the formation and utilization of protonassisted multi-electrons pathway in the photocatalytic CO 2 reduction process [47]. The Nanomaterials 2022, 12, 2284 9 of 15 CH 4 , CO and O 2 yields for CO 2 photoreduction reached 1.87, 0.0065 and 2.63 mmol/g cat , respectively. Compared with the photocatalytic activity of the reported catalysts (Table S3), the maximum rate of CH 4 product over Cu/Cu 2 O/WO 3 was obviously higher and high product selectivity was also obtained. To explore the influence of loading Cu on the photocatalytic performance, the controlled experiments in different atmospheres were further investigated (Figure 6b). In the CO/H 2 O atmosphere, more CH 4 molecules were produced over Cu/Cu 2 O/WO 3 compared with that of Cu 2 O/WO 3 . Meanwhile, in the N 2 /H 2 O atmosphere, the H 2 yield from water splitting was improved with the existence of Cu in the composite, which indicated that Cu nanoparticles promoted the electron transfer and proton aggregation on the catalyst surface, improving the hydrogenated process for CO 2 reduction over the Cu/Cu 2 O/WO 3 catalyst. The yields of products for the Cu/Cu 2 O/WO 3 composite were further investigated in the CO 2 /H 2 O atmosphere within 24 h of irradiation. As shown in Figure 6c, the total yields of CH 4 , H 2 and O 2 were enhanced with prolonging the irradiation time. The maximal yield rate of H 2 reached 0.013 mmol/g cat /h at 24 h, and those of CH 4 and O 2 at 18 h were found to be 0.104 and 0.147 mmol/g cat /h, respectively. The decreased yields of the photocatalytic CO 2 reduction were possibly due to oxidation of the formed carbonous compounds on the photocatalyst surface or the coverage of active sites by intermediates. Reproducibility and durability are critical for the long-term use of a catalyst in practical application. The results in Figure 6d show that the yields of CH 4 and H 2 for the typical Cu/Cu 2 O/WO 3 catalyst slightly decreased, and the O 2 yield obviously fell down after 5 cycles. Notably, after the 5th cycling experiment, the photocatalyst was regenerated by the electro-reduction method at 0.05 V in 0.5 M Na 2 SO 4 solution. In the 6th cycling experiment, all the yields were obviously promoted, which were similar to those of the fresh catalyst. To illustrate the stability of the Cu/Cu 2 O/WO 3 catalyst, XPS measurements were conducted to investigate the change in chemical composition after the 5th cycle. As shown in the Cu 2p and W 4f XPS spectra ( Figure S4), Cu (II) and W (V) were identified, which were probably derived from the oxidation of Cu (I) and reduction in W (VI), and W (V) possibly became the high recombination center of the photogenerated electrons and holes, leading to a slight decrease in the photocatalytic activity for CO 2 reduction. spectra ( Figure S4), Cu (II) and W (V) were identified, which were probably derived from the oxidation of Cu (I) and reduction in W (VI), and W (V) possibly became the high recombination center of the photogenerated electrons and holes, leading to a slight decrease in the photocatalytic activity for CO2 reduction.

Possible Photocatalytic Mechanism
Transient photocurrent response and photoluminescence spectra were measured to investigate the charge separation. Based on the photo-electric properties during several on-off illumination cycles (Figure 7a), the photocurrent density of Cu2O/WO3 was higher than those of the bare WO3 and Cu2O samples, and the Cu/Cu2O/WO3 sample exhibited the highest photo-induced density among them. Moreover, the photoluminescence spectra of the photocatalysts at the excitation wavelength of 405 nm are present in Figure 7b, and the emission intensity of the Cu/Cu2O/WO3 composite distinctly decreased compared with those of the WO3, Cu2O and Cu2O/WO3 samples. The above results indicate the promoted separation of photogenerated electron-hole pairs in the Cu/Cu2O/WO3 composite.

Possible Photocatalytic Mechanism
Transient photocurrent response and photoluminescence spectra were measured to investigate the charge separation. Based on the photo-electric properties during several onoff illumination cycles (Figure 7a), the photocurrent density of Cu 2 O/WO 3 was higher than those of the bare WO 3 and Cu 2 O samples, and the Cu/Cu 2 O/WO 3 sample exhibited the highest photo-induced density among them. Moreover, the photoluminescence spectra of the photocatalysts at the excitation wavelength of 405 nm are present in Figure 7b The band edge position and charge transport mode in the composite directly influenced the separation and redox ability of photoexcited charge carriers. Ultraviolet photoelectron spectra combined with Eg analysis were used to determine the electronic band structure of the catalyst. Figure 8a shows the UPS spectra of Cu2O and WO3. On the basis of the linear intersection method [48], the valence band (VB) of WO3 was estimated to be −7.33 eV (vs. vacuum), and the conduction band (CB) was −4.47 eV (vs. vacuum) based on the Eg value of WO3. According to the connection between vacuum energy and normal electrode potential (NHE) [29], the corresponding VB and CB positions of WO3 were 2.89 and 0.03 eV (vs. NHE), respectively. Similarly, the VB and CB values of Cu2O were separately calculated to be 0.91 and −1.14 eV (vs. NHE), which agreed well with previously reported results [31,49]. Hence, the heterostructure with staggered band alignment was successfully formed in this Cu2O/WO3 system, assuming that the possible band bending of the semiconductor was neglected. Based on the band energy structure of WO3 and Cu2O, two possible mechanisms for photo-induced carriers were proposed and described in Figure 8b,c. If the photo-excited charge carriers transferred according to the traditional model (Figure 8b), the photo-excited holes in the VB of WO3 would transfer to the VB of Cu2O, and the photo-excited electrons in the CB of Cu2O would migrate to the CB of WO3, where the accumulated electrons would not reduce CO2 to CO/CH4 or produce H2 from H2O on account of the more positive CB edge potential (0.21 eV vs. NHE) than the standard CO2 reduction potential [4]. Similarly, the accumulated holes in Cu2O would not accomplish H2O oxidation due to the more negative VB of Cu2O. Hence, the hypothesis of the traditional double-transfer model was invalid. The S-scheme charge transfer mode (Figure 8c) was proposed according to the enhanced photocatalytic activity of Cu/Cu2O/WO3. To verify the S-scheme charge transfer mode in this heterojunction, the XPS spectra of Cu2O/WO3 were measured in light and the results were shown in Figure  8d,e. The CB of Cu2O and WO3 were composed of Cu and W orbitals, respectively. Four peaks of W 4f in light shifted to a higher binding energy compared with these in the dark. Simultaneously, two peaks of Cu 2p in light reversely shifted. It was implied that the The band edge position and charge transport mode in the composite directly influenced the separation and redox ability of photoexcited charge carriers. Ultraviolet photoelectron spectra combined with E g analysis were used to determine the electronic band structure of the catalyst. Figure 8a shows the UPS spectra of Cu 2 O and WO 3 . On the basis of the linear intersection method [48], the valence band (VB) of WO 3 was estimated to be −7.33 eV (vs. vacuum), and the conduction band (CB) was −4.47 eV (vs. vacuum) based on the E g value of WO 3 . According to the connection between vacuum energy and normal electrode potential (NHE) [29], the corresponding VB and CB positions of WO 3 were 2.89 and 0.03 eV (vs. NHE), respectively. Similarly, the VB and CB values of Cu 2 O were separately calculated to be 0.91 and −1.14 eV (vs. NHE), which agreed well with previously reported results [31,49]. Hence, the heterostructure with staggered band alignment was successfully formed in this Cu 2 O/WO 3 system, assuming that the possible band bending of the semiconductor was neglected. Based on the band energy structure of WO 3 and Cu 2 O, two possible mechanisms for photo-induced carriers were proposed and described in Figure 8b,c. If the photo-excited charge carriers transferred according to the traditional model (Figure 8b), the photo-excited holes in the VB of WO 3 would transfer to the VB of Cu 2 O, and the photo-excited electrons in the CB of Cu 2 O would migrate to the CB of WO 3 , where the accumulated electrons would not reduce CO 2 to CO/CH 4 or produce H 2 from H 2 O on account of the more positive CB edge potential (0.21 eV vs. NHE) than the standard CO 2 reduction potential [4]. Similarly, the accumulated holes in Cu 2 O would not accomplish H 2 O oxidation due to the more negative VB of Cu 2 O. Hence, the hypothesis of the traditional double-transfer model was invalid. The S-scheme charge transfer mode (Figure 8c) was proposed according to the enhanced photocatalytic activity of Cu/Cu 2 O/WO 3 . To verify the S-scheme charge transfer mode in this heterojunction, the XPS spectra of Cu 2 O/WO 3 were measured in light and the results were shown in Figure 8d,e. The CB of Cu 2 O and WO 3 were composed of Cu and W orbitals, respectively. Four peaks of W 4f in light shifted to a higher binding energy compared with these in the dark. Simultaneously, two peaks of Cu 2p in light reversely shifted. It was implied that the photo-induced electrons transferred from WO 3 component to Cu 2 O component in this heterojunction [50,51]. Specifically, the photoexcited holes in the VB of Cu 2 O would transfer to WO 3 and recombine with the photoexcited electrons in the CB of WO 3 . The CB potential of Cu 2 O and VB potential of WO 3 thermodynamically realized photocatalytic CO 2 reduction and H 2 O oxidation, respectively, and the metallic Cu co-catalyst facilitated the reduction process dynamically.  [50,51]. Specifically, the photoexcited holes in the VB of Cu2O would transfer to WO3 and recombine with the photoexcited electrons in the CB of WO3. The CB potential of Cu2O and VB potential of WO3 thermodynamically realized photocatalytic CO2 reduction and H2O oxidation, respectively, and the metallic Cu co-catalyst facilitated the reduction process dynamically.

Conclusions
In summary, Cu and Cu/Cu2O species were synthesized on the hexagonal WO3 nanoflake films by the one-step electrodeposition method for photocatalytic CO2 reduction with water vapor. The obtained Cu/Cu2O/WO3 catalyst exhibited excellent photocatalytic performance under visible light irradiation (λ > 400 nm) due to the construction of heterojunction, and the CH4, H2 and O2 yields reached 2.43, 0.32 and 3.45 mmol/gcat after 24 h of illumination, respectively. Notably, CH4 molecules were generated as the major product over the Cu/Cu2O/WO3 catalyst, whereas Cu2O/WO3 facilitated CO generation. Efficient CH4 formation for the Cu/Cu2O/WO3 catalyst was attributed to the modification of Cu nanoparticles favoring electron-proton transfer from CO to CH4. The decreased photocatalytic activity in the cycling experiment was recovered by the regenerated treatment via electro reduction, removing the superfluous W(V) in the composite. Additionally, the S-scheme charge transfer mode and potential mechanism of CO2 reduction were proposed by the results of XPS measurement and photocatalytic performance under light illumination with a specific wavelength. The present research may provide a promising strategy to design ternary nanocomposite VLD photocatalysts and inspire further interest in tuning product selectivity for photocatalytic CO2 conversion.

Conclusions
In summary, Cu and Cu/Cu 2 O species were synthesized on the hexagonal WO 3 nanoflake films by the one-step electrodeposition method for photocatalytic CO 2 reduction with water vapor. The obtained Cu/Cu 2 O/WO 3 catalyst exhibited excellent photocatalytic performance under visible light irradiation (λ > 400 nm) due to the construction of heterojunction, and the CH 4 , H 2 and O 2 yields reached 2.43, 0.32 and 3.45 mmol/g cat after 24 h of illumination, respectively. Notably, CH 4 molecules were generated as the major product over the Cu/Cu 2 O/WO 3 catalyst, whereas Cu 2 O/WO 3 facilitated CO generation. Efficient CH 4 formation for the Cu/Cu 2 O/WO 3 catalyst was attributed to the modification of Cu nanoparticles favoring electron-proton transfer from CO to CH 4 . The decreased photocatalytic activity in the cycling experiment was recovered by the regenerated treatment via electro reduction, removing the superfluous W(V) in the composite. Additionally, the S-scheme charge transfer mode and potential mechanism of CO 2 reduction were proposed by the results of XPS measurement and photocatalytic performance under light illumination with a specific wavelength. The present research may provide a promising strategy to design ternary nanocomposite VLD photocatalysts and inspire further interest in tuning product selectivity for photocatalytic CO 2 conversion.