The polycrystalline InTaO₄ was synthesized by a solid-state reaction method as reported in literature [24]. The pre-dried In₂O₃ and Ta₂O₅ were used as the starting materials. The stoichiometric amounts of precursors were mixed and reacted in an aluminum crucible in air at 1100˚C for 12 h. The material was stirred at least 3 times during preparation to ensure well mix of starting materials.

Co₃O₄-InTaO₄ samples with various Co₃O₄ cocatalyst (0.3 wt%, 0.5 wt% and 1 wt%, respectively) were prepared by incipient-wetness impregnation with aqueous solution of Co(NO₃)₂. After preparation, the sample was heated by a water bath at 100˚C. The dried powder was then calcined at 400˚C for 4 h in an oven. The sample was pretreated by H₂ (200 Torr) reduction at 500˚C for 2 h and subsequent O₂ (100 Torr) oxidation at 200˚C for 1 h. The core-shell structure of metallic Co and Co₃O₄ was formed by this reduction-oxidation procedure.

Photocatalytic reactions were carried out in a continuous flow reactor. The detailed reaction procedure was described in previous literature [25]. The catalyst powder ( 0.14 g ) was dispersed in a reactant solution (50 mL) in a down-window type irradiation cell made of Pyrex glass (75 ml). 0.2 M Sodium hydroxide aqueous solution or 0.2 M potassium bicarbonate aqueous solutions were employed as an absorbent of carbon dioxide and the ultra pure CO₂ were added continuously into the reactor for 1 h to remove the oxygen in the water, and saturated carbon dioxide in the solution. Using the cooling system combined with water pump, the temperature of the reactor was maintained at room temperature. Light on to start the reaction, and the irradiation was continued for 20 h. The light source was a 500 W halogen lamp (Ever bright; H-500). After reaction for 20 h, the reaction solution was centrifuged for 10 min to separate the reaction products from the powder catalyst. 10 mL of the upper stratum was taken for analyzing the concentration of methanol. The amount of methanol was determined by a gas chromatography equipped with a flame ionization detector, using Poropack-QS column.

Figure 1 shows the XRD patterns of the Co₃O₄-InTaO₄ samples. The XRD patterns of InTaO₄ samples are well consistent with monoclinic InTaO₄ pattern and space group P2/c, indicating that the samples were fully crystallized in the wolframite-type structure. InTaO₄ has major peaks at around 2θ = 23.967˚ (−110),

29.356˚ (−111) and 29.899˚ (111) [20] [21] [22] [23] [24]. The lattice parameters of the crystal were refined as: a = 4.83300 (−1) Å, b = 5.77800 (1) Å, c = 5.15700 (1) Å and β = 91.380˚. The indexed results are in good agreement with those reported in the JCPDS database card (No. 25-0391). Zou and his coworkers [20] have reported the structural refinements of InTaO₄. The InTaO₄ compound belongs to a monoclinic system with space group P2/c, a = 5.1552 (1), b = 5.7751 (1), c = 4.8264 (1) Å and β = 91.373 (1)˚. The structure is composed of two kinds of octahedral: InO₆ octahedron and TaO₆ octahedron. The InO₆ octahedron connects to each other to form zigzag chains by sharing edge. These InO₆ chains are connected through TaO₆ octahedron to form the three dimensional network.

Figure 1 also shows that the characteristic XRD peaks corresponding to Co species such as Co₃O₄ were not observed in the XRD patterns, indicating that the Co species on InTaO₄ was too small to detect. There was no difference in XRD patterns between InTaO₄ and 1.0 wt% Co₃O₄-InTaO₄, indicating that the addition of Co₃O₄ cocatalyst on the surface of InTaO₄ did not change the bulk structure of InTaO₄. However, it could modify the surface of InTaO₄ as discussed in the latter section. One can conclude that Co₃O₄ nanoparticles were well dispersed on the surface of InTaO₄.

Figure 2 shows the SEM photographs of InTaO₄ samples with various amounts of Co₃O₄ cocatalysts. The particle size of InTaO₄ was about 0.5 μm. Many nano Co₃O₄ particles were present on InTaO₄ surface, in consistent with XRD results.

Photocatalytic activity is dependent on the band structure of semiconductor. The information of band structure is very important for understanding photocatalytic reaction. Figure 3 shows the diffuse reflectance spectra of InTaO₄ samples with various amounts of Co₃O₄ loading. It shows higher light absorption ability of Co₃O₄/InTaO₄ in the visible light compared with InTaO₄. The band gap of InTaO₄ was 3.0 eV. For 0.3, 0.5, and 1.0 wt% Co₃O₄-InTaO₄ catalysts after calcinations, an obvious absorption in the visible light region were observed on all catalysts. The absorbance of the sample increased with increasing the amount of Co₃O₄ cocatalysts. The band gap of 0.3, 0.5, and 1.0 wt% Co₃O₄-InTaO₄ were calculated to be 2.8 eV, 2.7 eV and 2.6 eV, respectively (Table 1). The results indicate that adding Co₃O₄ cocatalyst on InTaO₄ changed the band gap. The bangap of the catalyst decreased with an increase of Co₃O₄ loading.

The activities of carbon dioxide reduction on InTaO₄ samples with various Co₃O₄ loadings are shown in Figure 4. All catalysts produced methanol from the photoreduction of CO₂ under visible light irradiation. No other products were detected in gas phase and liquid phase. The rate of the reaction product increased linearly with the visible light-irradiation time, and the reaction stopped immediately when the irradiation was ceased. The formation of the reaction product was not detected under dark conditions. The reaction rate varied with the smount of cocatalyst. The results in Figure 4 were obtained from the InTaO₄ catalyst with Co₃O₄ cocatalyst suspended in 0.2 M NaOH and 0.2 M KHCO₃ aqueous solution. The highest methanol yield of 1.0 wt% Co₃O₄-InTaO₄ was 1.150 μmol.h⁻¹ g catal.⁻¹. In the NaOH solution, 0.5 wt% Co₃O₄-InTaO₄ demonstrated the highest methanol yield, and the Co₃O₄ cocatalyst enhanced the production of methanol. The Co₃O₄ cocatalyst not only provides reaction centers, which effectively transfer the electrons from the surface of catalysts to Co, but also enhances the light absorbance.

The results showed that the photocatalytic reduction was induced by the visible light irradiation. The formation rate of methanol increased with the presence

of cocatalysts on InTaO₄ photocatalysts. The photocatalyst had a higher activity in KHCO₃ aqueous solution than in NaOH solution, in agreement with literature results [17]. The InTaO₄ catalyst with 1.0 wt% Co₃O₄ cocatalyst in KHCO₃ aqueous solution gave the highest yield of methanol among all catalysts.

In the case of 1.0 wt% Co₃O₄-InTaO₄, Co₃O₄ species were loaded on InTaO₄ as nanoparticles, which were not observed by SEM analysis. However, after pretreatment, there was a formation of bulky Co₃O₄ particles on InTaO₄ due to aggregation of Co₃O₄ nanoparticles, leading to low photocatalytic activity. Bulk Co₃O₄ is a p-type semiconductor, which induces the formation of positive holes. For photoreduction of CO₂, hole scavengers are necessary to facilitate photoreduction; consequently, bulk Co₃O₄ reduced the photocatalytic activity of 1.0 wt% Co₃O₄–InTaO₄ catalyst. Hence, it is necessary to avoid the formation of bulk Co₃O₄.

The characterization results of 0.5 wt% Co₃O₄-loaded InTaO₄ photocatalyst showed the presence of ultra-fine Co₃O₄ thin films on metallic cobalt particles. The high dispersion of Co₃O₄ particles on the surface and interface of InTaO₄ plays a major role in determining its photocatalytic activity. The metallic cobalt acts as a bias for electron transfer from InTaO₄ to Co₃O₄ layer and the excited electron can migrate easily to the surface to facilitate photoreduction of CO₂. Methanol acts as a hole scavenger to improve the yield.

It should be noted that for dry InTaO₄, all the donor states are occupied and no optical transitions from the valence band to the donor state occurs. Instead, when InTaO₄ is immersed in water, partial depletion of the donor states will occur. This leads to band bending and the formation of a depletion layer, as reported in literature [23] [27] [28] [29]. The ionized donor states can be filled through optical excitation of valence band electrons, which explains the sub-bandgap optical absorption of InTaO₄ and high photoactivity of InTaO₄ in liquid phase reduction of CO₂. Zou et al. [20] reported that the bottom of conduction band of Ta_3d is lower than conduction band level of Co₃O₄. Accordingly, the conduction band level of the Co₃O₄/InTaO₄ is not high enough for electrons transfer across InTaO₄ and Co₃O₄ interface. Co cations are presumably located on the Ta³⁺ sites, especially when one considers that the formation of singly charged acceptor defects is energetically much more favorable than the formation of the triply charged defects that would be formed if any Co would substitute for Ta⁵⁺. Since Co₃O₄ was added after formation of full crystallite of InTa₄. Co₃O₄ did not incorporate into bulk InTaO₄ crystal. Co₃O₄ was incorporate into the surface structure of InTaO₄ as Co_xInTaO_4-x. It resulted in more defect sites on the surface of InTaO₄ and changed the valence band structure and the surface became a Schottky barrier to suppress the recombination of electron and holes. The higher light absorption ability of Co₃O₄/InTaO₄ in the visible light compared with InTaO₄ was also responsible for the high activity of Co₃O₄/inTaO₄ catalysts.