In Situ Formation of Z-Scheme Bi₂WO₆/WO₃ Heterojunctions for Gas-Phase CO₂ Photoreduction with H₂O by Photohydrothermal Treatment

Abstract

We report a new photohydrothermal method to prepare a Bi₂WO₆/WO₃ catalytic material for CO₂ photoreduction by solar concentrators. The photohydrothermal treatment improves the physico-chemical properties of the Bi₂WO₆/WO₃ material and forms well contact Bi₂WO₆/WO₃ heterojunctions, which increase the maximum reaction rate of CO₂ photoreduction to 8.2 times under the simulated light, and the hydrocarbon yield under the real concentrating solar light achieves thousands of μmol·g_cata⁻¹. The reason for the high activity is attributed to the direct Z-scheme effect of Bi₂WO₆/WO₃ heterojunctions and the photothermal effect during the course. These findings highlight the utilization of solar energy in CO₂ photoreduction and open avenues for the rational design of highly efficient photocatalysts.

1. Introduction

Solar energy is a massive, free, and non-polluting renewable energy source. Increasing the utilization efficiency of solar energy is one of the best solutions for a sustainable society. To facilitate storage and terminal utilization, it is better to convert solar energy into other forms, such as electricity or chemical compounds [1,2,3]. Photocatalytic reduction of CO₂ with H₂O, as occurs in green plants, can generate platform compounds with abundant energy such as CH₄ and methanol, and thus forms one of the main routes for solar energy transformation [4,5,6].

CO₂ photoreduction depends on the energy input of solar light and the function of the photocatalyst. CO₂ photoreduction entails the adsorption of photons in the incident light to generate electron-hole pairs, the separation/migration of photon-generated charge carriers, and the surface reaction of charge carriers with reactants [7,8,9]. Therefore, a desirable photocatalyst has high light harvesting efficiency, charge separation efficiency, charge migration and transport efficiency, and charge utilization efficiency for photocatalysis, and the promotion of any step is beneficial for the general efficiency. Tremendous efforts have been devoted to the development of effective photocatalysts [10]. However, at present, the general solar energy conversion efficiency remains at a poor level. This is because CO₂ molecules are stable, which means that only a small part of shortwave high-energy photoelectrons in solar light can activate them. Secondly, there are many reaction steps from CO₂ to hydrocarbon, while the recombination of the photo-generated electron-hole pair only needs one step, which is unfavorable to CO₂ reduction [11,12]. Most solar energy is thus converted into useless low-grade heat, which severely hinders the progress of CO₂ photoreduction.

The recombination of the photo-generated electron-hole pair also threatens the process intensification of CO₂ photoreduction, as the higher the incident light intensity, the more recombination of the pairs. Therefore, only few researchers have conducted studies in this domain. Rossetti et al. [13] proposed a concept of a high-pressure photoreactor that can operate under pressure up to 20 bar. Wu et al. [14] increased some incident light of a fiber reactor with a spherical solar concentrator. With an additional step, solar energy can generate high-grade heat by large solar concentrators. The high-grade heat then can couple with the light to yield a photothermal approach, i.e., a catalytic process driven by the photochemical and thermochemical forces together [15].

The photothermal approach also broadens the absorption of the solar spectrum and provides a competitive way of raising the efficiency of solar energy transformation. The high incident light intensity by concentrating technology will contribute to a considerable increase in the CO₂ reaction rate. With the same catalyst, the CO₂ reaction rate under concentrating conditions can reach hundreds of times the rate under non-concentrating situations [16,17].

More interestingly, the photothermal effect is extended to the catalytic materials’ preparation process. It is known that high-temperature hydrothermal treatment is a useful method to improve the properties of different materials, such as TiO₂ nanotubes, zeolites, etc. [18,19]. The high temperature and pressure can change the physical properties of the crystallized anatase powder, in turn improving the subsequent phase change properties of the anatase/rutile phase change. Similarly, photohydrothermal treatment, i.e., an environment with high light intensity, temperature, and pressure conditions, is expected to evolve the catalytic materials into a new, profitable state.

Bi₂WO₆ (band gap 2.8 eV) is one of the most studied catalysts in CO₂ photoreduction [20,21,22,23,24,25,26]. It has been demonstrated to be an active photocatalyst in the visible light band, and the formation of heterojunctions of Bi₂WO₆ with some other oxides, such as TiO₂ or WO₃, can further improve the activity [27,28]. However, it is necessary to find a new way to realize the well contact of two oxide phases and obtain adequate heterojunctions. In this manuscript, we present a new photohydrothermal method to promote the formation of Bi₂WO₆/WO₃ material with heterojunctions. The CO₂ photoreduction tests show that the Bi₂WO₆/WO₃ catalyst has favorable photocatalytic activity under simulated and real solar light. Furthermore, we propose the possible mechanism of CO₂ photoreduction in the reaction process. The photohydrothermal route can be a novel green technology for the preparation of similar materials and broaden the solar energy utilization scope.

2. Results and Discussions

2.1. CO₂ Photoreduction Performance of Bi₂WO₆/WO₃ under Real and Simulated Light

Figure 1 displays the yield and distribution of CO₂ photoreduction products on photohydrothermally treated Bi₂WO₆/WO₃ materials driven by real solar light. At a concentration ratio (CR, the ratio of incident light area to the catalyst disc area) of 1, i.e., natural solar light without concentration, only CH₄ is detected, and the yield is about 2.57 μmol·g_cata⁻¹·h⁻¹ after 5 h of reaction. The yield is lower than that under simulated light from a 300 W Xe lamp, which is about 3.43 μmol·g_cata⁻¹·h⁻¹. Then, when the CR increases to 400, in addition to CH₄, two more products appear: C₂H₄ and C₂H₆. After 5 h, the average yield rate of CH₄ is about 166.13 μmol·g_cata⁻¹·h⁻¹, that of C₂H₄ is 56.42 μmol·g_cata⁻¹·h⁻¹, and that of C₂H₆ is 28.11 μmol·g_cata⁻¹·h⁻¹. The total CO₂ conversion reaches 125.33 μmol and 1.12% in the reactor. The average rate increases to CH₄ 304.94 μmol·g_cata⁻¹·h⁻¹, C₂H₄ 62.70 μmol·g_cata⁻¹·h⁻¹, and C₂H₆ 54.66 μmol·g_cata⁻¹·h⁻¹, and the total CO₂ conversion reaches 211.15 μmol and 1.89% (

Table 1. STC average efficiencies of Bi₂WO₆/WO₃ under concentrating solar light.

CR	Yield/μmol·g−1			STCaverage/%		CO2 Conversion/%
	CH4	C2H4	C2H6	UV	Total
400	830.67	282.09	140.54	0.42	0.03	1.12
600	1524.70	313.52	273.31	0.72	0.05	1.89
800	368.95	83.05	84.05	0.28	0.02	0.63

The STC is calculated based on the total incident light intensity in Hangzhou, 58 mW/cm², and the UV part is considered to be 7% of the total light.

and

Table 2. STC max efficiencies of Bi₂WO₆/WO₃ under concentrating solar light.

CR	Yield/μmol·g−1			STCmax/%
	CH4	C2H4	C2H6	UV	Total
400	198.31	152.42	43.33	0.84	0.06
600	430.18	127.79	288.29	1.73	0.12
800	117.50	30.42	22.73	0.28	0.02

) at CR 600. However, all the yield rates decrease when the CR continually increases to 800.

The transformation efficiency of solar energy to chemical (STC) is calculated according to the formula:

$$\mathrm{STC}=\frac{\mathrm{Output}\mathrm{energy}\mathrm{as}\mathrm{a}\mathrm{chemical}}{\mathrm{Energy}\mathrm{of}\mathrm{incident}\mathrm{solar}\mathrm{light}}=\frac{\mathrm{r}\times \Delta {\mathrm{G}}_{\mathrm{r}}}{{\mathrm{P}}_{\mathrm{sun}}\times \mathrm{S}}$$

(1)

where r is the product yield rate, $\Delta {\mathrm{G}}_{\mathrm{r}}$ is the Gibbs free energy, and ${\mathrm{P}}_{\mathrm{sun}}$ and S are the incident light intensity and incident light area, respectively. From Table 1 and Table 2, it can be seen that the maximum STC can reach 0.12% in the total solar light spectrum.

A large concentration ratio can raise the reaction temperature, which will give the impression that the high reaction rate is from the reaction conditions and not from the catalyst properties. Therefore, the yield and distribution of CO₂ photoreduction products on the Bi₂WO₆/WO₃ catalysts treated by the photohydrothermal method with different times were tested using a simulated light (a 300 W Xe lamp) reactor system, as shown in Figure 2. On Bi₂WO₆/WO₃ without photohydrothermal treatment, only CH₄ is detected, and the yield is about 17.14 μmol·g_cata⁻¹ after 5 h of reaction. The yield is slightly larger than that under the natural light in Figure 1. Then, when it is photohydrothermally treated at CR 600 for 3 h, C₂H₄ appears. After a 3 h reaction, the yield of CH₄ reaches 101.44 μmol·g_cata⁻¹ and that of C₂H₄ reaches 27.10 μmol·g_cata⁻¹. The total CO₂ conversion reaches 128.54 μmol·g_cata⁻¹. When the photohydrothermal treatment time prolongs to 5 h, both C₂H₄ and C₂H₆ appear again. After a 5 h reaction, the yield of CH₄ reaches 143.38 μmol·g_cata⁻¹. The total CO₂ conversion reaches 204.26 μmol·g_cata⁻¹.

The results in Figure 1 and Figure 2 illustrate that the catalytic activity of Bi₂WO₆/WO₃ material is improved after the photohydrothermal treatment. The appearance of ethylene and ethane also enriches the types of products, which means that it is possible to directly obtain C₂+ products by this route. The results are also reasonable, as the CO₂ photoreduction reaction is similar to the CO₂ hydrogenation reaction. In the reaction sequence of CO₂ reduction with H₂O, the H₂O first dissociates into H₂ and O₂, and then H₂ reacts with CO₂ [29]. The later reaction is known to be able to obtain molecules with multiple carbon atoms, such as ethanol, ethene, and even higher hydrocarbons [30]. High-carbon products are not often discussed, perhaps because the yield of these products is too low to be detected. Here, however, the high yield discloses their existence. The high yield proves that Bi₂WO₆/WO₃ is an excellent catalytic material for CO₂ photoreduction. The results in Figure 1 also illustrate that Bi₂WO₆/WO₃ possesses high CO₂ photoreduction activity under real concentrating solar light. The CO₂ reaction rates at CR 400, 600, and 800 are several hundred times greater than the rate under natural light. It is known from the photocatalytic reaction kinetics that the order of light intensity in photocatalysis decreases with the incident light strength, i.e., the light utilization efficiency will decrease with the incident light strength, and the increment in incident light intensity cannot induce the corresponding increment in the reaction rate. Therefore, the CO₂ reaction rate here indicates that there might be another factor. The measured high temperatures in the reactor under high concentration ratios also indicate that the thermal effect is favorable for CO₂ photoreduction. In general, the CO₂ photoreduction results of Bi₂WO₆/WO₃ driven by the real and simulated solar light illustrate that there is a photothermal, even a photohydrothermal, effect yielded by the concentrating solar light technology, which is beneficial for the whole process.

2.2. Characterization of Bi₂WO₆ and Bi₂WO₆/WO₃ Samples

The texture properties of the Bi₂WO₆/WO₃ samples are listed in

Table 3. Texture properties of the Bi₂WO₆/WO₃ samples before and after photohydrothermal treatment.

Sample	Surface Area (m2/g)	Pore Volume (cc/g)	Average Pore Diameter (nm)
Bi2WO6/WO3	25.8	0.101	14.7
Bi2WO6/WO3 (CR 400)	26.3	0.118	17.9
Bi2WO6/WO3 (CR600)	17.0	0.217	51.2
Bi2WO6/WO3 (CR800)	5.6	0.119	85.2

. The specific surface area of the sample does not change considerably after being treated at CR 400, while it noticeably decreases after being treated at CR 600 and 800. In addition, the average pore diameter increases with CR 600 and 800 treatment, which indicates the size growth in size of the nanoparticles.

The crystal structures of the pure Bi₂WO₆ and Bi₂WO₆/WO₃ samples were detected by XRD technology. The results are shown in Figure 3. The characteristic peaks of Russellite Bi₂WO₆ (JCPDS no. 26-1044) and the monoclinic WO₃ (JCPDS no. 43-1035) can be identified by the patterns. After photohydrothermal treatment, the diffraction peaks of the samples changed. As seen in Figure 3, the peaks of WO₃ at values of 23.2°, 23.6°, and 24.4° corresponding to (002), (020), and (200) shift slightly and transform into two diffraction peaks corresponding to (020) and (200). [31] The feature peaks of Bi₂WO₆ at values of 28.6° and 33.03° become clearer.

The morphology and microstructure of Bi₂WO₆/WO₃ before and after photohydrothermal treatment were investigated using SEM and TEM. Typical cuboids with a moderate size were found in the parent Bi₂WO₆/WO₃. The photohydrothermally treated Bi₂WO₆/WO₃ displays the size growth of the particles, which is also seen in TEM (Figure 4). From the TEM image of Bi₂WO₆/WO₃ heterojunction (Figure 5), the lattice fringes of the sample after photohydrothermal treatment are clearer than those of the parent sample. Two oxide phases of Bi₂WO₆ and WO₃ are more clearly observed and closely contact to form an intimate interface after photohydrothermal treatment. The value of 0.248 nm corresponds to the (202) crystallographic plane of the orthorhombic Bi₂WO₆ crystal, which is also in accordance with the XRD results in Figure 3. The value of 0.342 nm corresponds to the (012) crystallographic plane of the WO₃. This result suggests that the photohydrothermal treatment could improve the Bi₂WO₆/WO₃ heterojunctions in the structure. The tight coupling is favorable for the charge transfer between WO₃ and Bi₂WO₆ and promotes the separation of photogenerated electron-hole pairs, subsequently improving the photocatalytic activity.

The composition and valance state of Bi₂WO₆/WO₃ are demonstrated via XPS technology, and the spectra are shown in Figure 6. The XPS spectra confirm that Bi, W, and O elements coexist in the samples. Figure 6a shows that the typical peak of O 1s, located at around 530 eV, can be deconvoluted into two peaks from the lattice Bi–O and W–O. After photohydrothermal treatment, a new peak appears from O–H or oxygen vacancy. The characteristic peaks of W 4f (Figure 6) are located at 37.2 and 35.1 eV, which conform to W 4f_5/2 and 4f_7/2, respectively, revealing that W presents with W⁶⁺. Figure 6c shows that two peaks at 164.2 and 158.9 eV exist on the XPS spectrum of Bi, which are related to Bi 4f_5/2 and Bi 4f_7/2, respectively. The phenomenon demonstrates that Bi exists in the photocatalyst with Bi³⁺ [32,33]. After the treatment, the two peaks slightly move toward a lower altitude, indicating the interaction between Bi₂WO₆ and WO₃.

2.3. Discussions

The catalytic performance of Bi₂WO₆/WO₃ under real and simulated solar light proves two points:

1. The photohydrothermal treatment improves the catalytic performance of Bi₂WO₆/WO₃ under real light and 300 W Xe light.

2. The catalytic performance of Bi₂WO₆/WO₃ under simulated solar light from a 300 W Xe source further confirms that the effect is part of the improvement in catalyst properties, as the reaction conditions provided by Xe light are similar to the classic CO₂ photoreduction reaction conditions, which exclude the possible influence from the reaction conditions provided by real concentrating solar light.

We then analyzed how the photohydrothermal treatment influences Bi₂WO₆/WO₃. By the characterizations, it can be seen that the photohydrothermal treatment decreases the specific surface area (BET), decreases the crystallinity of Bi₂WO₆ or WO₃ (XRD), and increases the particle sizes (SEM). All of these changes do not appear beneficial for the catalytic activity. Considering the characterization and our initial intention, it is reasonable to accept that the formation of the Bi₂WO₆/WO₃ heterojunctions (TEM and XPS) is a possible reason for such activity. To increase the likelihood of the hypothesis, the activity of the pure Bi₂WO₆ and WO₃ with real light as the source was tested, as shown in Figure 7 and Figure S2. It can be seen that pure Bi₂WO₆ and WO₃ are lower than Bi₂WO₆/WO₃ under the same conditions.

In general, the activity promotion of Bi₂WO₆/WO₃ does not stem from the specific surface area, as the specific surface area decreases. It also does not come from the crystallinity, as can be seen in Figure 3. The activity of the pure Bi₂WO₆ and WO₃ material with the same preparation procedure is proven to be lower than that of the Bi₂WO₆/WO₃ materials. That is to say, the interaction of Bi₂WO₆/WO₃ must have happened in some respect, and this is favorable for CO₂ photoreduction. The interaction between two species is often called the support effect in thermal catalysis, while in the photocatalyst field, it is often called a heterojunction. Although more evidence is needed, it appears reasonable to adopt the heterojunction theory to explain the results here [34,35].

In Bi₂WO₆-related materials, it is common to construct a heterojunction structure by adjusting the components of Bi or W elements. For example, Aranda-Aguirre et al. [36] prepared Bi₂O₃/Bi₂WO₆ thin films for the photo-electrocatalytic degradation of histamine. Chung et al. [37] constructed Bi₂WO₆ and WO₃ heterojunction photoanodes for improved charge transportation, and He et al. [38] synthesized a core/shell WO₃ (core)/Bi₂WO₆ (shell) structure. Gui et al. [39] achieved the heterojunction of Bi₂WO₆/WO₃ with a one-step hydrothermal method in 2012. In 2020, Chen et al. [40] further discussed the Bi₂WO₆/WO₃ heterojunction by facet engineering with salicylic acid removal reaction. Under visible light irradiation, Bi₂WO₆ and WO₃ are excited simultaneously, and electron-hole pairs are generated. For certain faces, i.e., WO₃(001) and (110)/Bi₂WO₆, the photogenerated electrons can instantly transfer from CB of WO₃ to VB of Bi₂WO₆, and then combine with the holes of Bi₂WO₆, leading to the high-efficiency carriers’ separation in the composite.

Based on the results and discussion above, a possible direct Z-scheme photocatalytic mechanism for the Bi₂WO₆/WO₃ heterojunctions in CO₂ photoreduction is proposed and schematically exhibited in Scheme 1. Under the incident light irradiation, both Bi₂WO₆ and WO₃ could be excited to generate electrons (e⁻) and holes (h⁺). Then, the photogenerated electrons of WO₃ will transfer to the valence band of Bi₂WO₆, leaving the electrons in Bi₂WO₆ and the holes in WO₃, resulting in electrons with higher reduction potential in Bi₂WO₆ and holes with higher oxidation potential in WO₃. The electrons in Bi₂WO₆ will react with the adsorbed CO₂ and reduce it to hydrocarbons. The accumulated holes in WO₃ will be consumed by oxidizing H₂O to O₂. Therefore, the Bi₂WO₆/WO₃ heterojunctions form a direct Z-scheme structure and show the efficient activity of photocatalytic CO₂ reduction and selectivity toward hydrocarbons.

Finally, we investigated if the effect of photohydrothermal treatment is better than the effect of the normal hydrothermal treatment. We tested the activity of Bi₂WO₆/WO₃ treated by the normal hydrothermal method, as shown in Figure 8. The activity of the normal hydrothermal method is lower than that of the photohydrothermal method. The light must play a role and induce the crystal growth. Thus, photohydrothermal treatment has certain merits.

3. Materials and Methods

The pure Bi₂WO₆ and parent Bi₂WO₆/WO₃ material were first prepared by a co-impregnation method. For the pure Bi₂WO₆, 0.9702 g of bismuth nitrate pentahydrate was dissolved in a 100 mL beaker with 20 mL of deionized water under stirring. When white precipitation appeared, the pH was adjusted until the white precipitation disappeared, and 2 g of sodium dodecyl sulfate surfactant was added to obtain solution A. To obtain solution B, 0.254 g of hydrate ammonium tungstate was added to another 100 mL beaker with 20 mL of water in a water bath at 80 °C under stirring. Solution B was then added to solution A dropwise with stirring to yield solution C. Then, 5 mL of ethylene glycol was added to solution C and its pH was adjusted to 7. After stirring solution C for 2 h, it was transferred to a clean autoclave and placed in an oven at 160 °C for 20 h. The sample was taken out and washed several times with deionized water and ethanol, dried in an 80 °C oven overnight, and finally calcined in a muffle at 500 °C for 2 h.

For the parent Bi₂WO₆/WO₃, excess 0.4 g WO₃ powder was added to solution A during the solution’s preparation step [41,42]. The photohydrothermal treatment of the Bi₂WO₆/WO₃ sample was carried out in a homemade concentrating solar reactor system. A detailed illustration of the system can be seen in the Supplementary Materials. A proper amount of the Bi₂WO₆/WO₃ sample was put into the reactor with a certain amount of liquid H₂O. The reactor was then sealed and purged with highly pure CO₂ gas for 30 min to remove the impure gases from the reactor. The reactor was fixed on the concentrating solar light system. By the detector and meter, the reactor angle was adjusted to let solar light enter the reactor to reach a certain light intensity, temperature, and pressure. The treatment was maintained for certain time (1–2 h).

The crystal structures of the pure Bi₂WO₆ and Bi₂WO₆/WO₃ materials before and after reaction were characterized by an X-ray diffractometer apparatus with Cu-Kα source (X’Pert PRO, PNAlytical, The Netherlands). The scanning angle was set from 10° to 80° (2θ) with a rate of 0.02°. The surface morphology picture of the Bi₂WO₆/WO₃ material was taken by an FESEM apparatus (field emission scanning electron microscopy, Hitachi S-4700, Hitachi Ltd., Chiyoda, Japan). The voltage range was 0.5–30 kV, and the acceleration voltage was set to 15 kV. The HRTEM (Tecnai G2 F30 S-Twin, FEI, The Netherlands) was used to distinguish the morphology change in the Bi₂WO₆/WO₃ materials before and after the reaction. The valence state of the Bi₂WO₆/WO₃ material was recorded by XPS spectra on a clutches spectrophotometer(Kratos AXIS Ultra DLD, Shimadzu, Kyoto, Japan ).

Experiments of the photocatalytic reduction of CO₂ reactivity were performed under a 300 W Xe lamp and natural light. Before the reaction began, the reactor was cleaned and dried. Deionized water was added, and CO₂ gas was used to check for leakage, remove impure gases, and act as a reactant. Sampling analysis was performed every 1 h. The light intensity was measured with a light intensity meter, and the temperature was recorded by a k-type thermocouple. Samples were analyzed by gas chromatography (GC-2014, Shimadzu, Kyoto, Japan) with an FID detector equipped with HT-PLOT Q capillary columns that can detect various hydrocarbon compounds from C1 to C6.

4. Conclusions

We report a new photohydrothermal method for photocatalyst preparation. The photohydrothermal treatment can reconstruct the morphology of the Bi₂WO₆/WO₃ material and form well contact Bi₂WO₆/WO₃ heterojunctions, showing high CO₂ photoreduction activity under the simulated and concentrating real solar light. The photothermal and photohydrothermal effects are expected to be beneficial not only for the CO₂ photoreduction reaction but also for the preparation of similar photocatalysts, and accelerate the research progress of solar fuels.