The crystal structure and the phase purity of the prepared samples were confirmed by X-ray diffraction (Shimadzu 6,000) with Cu-Kα radiation, λ = 1.5406 Å. The surface morphology was analyzed using a field emission scanning electron microscope (FESEM model—CARLZEISS SIGMA HV). The chemical composition analysis was studied by Thermo Scientific K-Alpha X-ray Photoelectron Spectroscopy (using ESCA-3400 XPS from Kratos Analytical, Shimazu). The optical band gap was estimated from UV-Vis diffused reflectance spectra using the JASCO V-670 double-beam spectrophotometer model. Synchrotron X-ray absorption spectra (Cu K-edge), including X-ray absorption near edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectra, were obtained at BL17C of the National Synchrotron Radiation Research Center (NSRRC), Taiwan. The specific surface area was calculated from N₂ adsorption-desorption isotherms measured using the ANTON PAAR NOVATOUCH LX2 (Graz, Austria), using the Brunauer–Emmett–Teller (BET) equation.

The photodegradation of MO dye was examined in the presence of the prepared CuTa₂O₆ photocatalyst. In a typical experiment, 50 mg of the photocatalyst was added to 100 mL of aqueous MO with an initial concentration of 10 mg/L. Before irradiation, the mixture was continuously stirred at room temperature in the dark for 30 min to achieve an adsorption/desorption equilibrium. Then, the mixed solution was irradiated with visible light using a 75 W high-pressure mercury vapor lamp. During the irradiation, about 2 mL of the suspension was taken at regular intervals (10 min) and analyzed using the UV-Vis spectrophotometer to determine the concentration of MO dye; the characteristic absorption of MO was about 464 nm (Mageshwari et al., 2013).

Figure 1 shows XRD patterns of the prepared CuTa₂O₆ at different annealing temperatures. The diffraction peak reveals the polycrystalline nature of the prepared CuTa₂O₆ phase. The sample annealed at 500°C exhibits the orthorhombic phase of CuTa₂O₆ and agrees with JCPDS card no: 87–0,357. Then, increasing the annealing temperature to 700°C led to the formation of a different phase. The observed diffraction peaks are associated with the cubic phase of CuTa₂O₆, which agrees with JCPDS card 70–0,611. Further increase in the annealing temperature to 900°C did not modify the phase formation. Therefore, increasing the annealing temperature significantly affects the phase formation of CuTa₂O₆. The intensity of the diffraction peaks is higher, and the full width at half maximum (FWHM) is also slightly increased with an increase in the annealing temperature. The average crystallite size was calculated using the Debye-Scherer formula D = K λ β cos ⁡ θ where K is the shape factor of the particles that are commonly called the Scherer constant (k = 0.89); λ is the wavelength of the X-ray radiation (λ = 0.1542 nm Cu-Kα radiation); θ is the Bragg angle, and β is the FWHM of the peak.

The difference in the XRD peak width is small between 700°C and 900°C. However, the actual variation is shown in Supplementary Figure S1 (supporting information), which evidences the increased value for full width at half maximum (FWHM). The intensity of the diffraction peaks and the FWHM) also increased as the temperature increased. The higher the FWHM, the smaller the crystallite according to the Debye-Scherer formula. Therefore, the calculated average crystallite size is found to be 32.28 nm, 29.79 nm, and 28.13 nm for the samples that were annealed at 500°C, 700°C, and 900°C respectively. An increase in the annealing temperature reduces the crystallite size, this could be due to the higher annealing temperature which might produce a relatively large temperature gradient within the submicrosheres, resulting in the formation of solid/liquid biphase core-shell structures through the non-equilibrium heating process (Tao et al., 2013). As heterogeneous nucleation only needs a low activation energy, the co-existence of heterogeneous phases can be helpful for the crystalline nucleation of CuTa₂O₆. Also, it is believed that the more generated nuclei, the smaller the crystallite size.

Figure 2A displays the nearly spherical shaped structure and uniform in size for the sample annealed at 500°C. It seems an increase in the annealing temperature to 700°C lead to agglomeration of these particles, as shown in Figure 2B. Further increase in the annealing temperature (900°C) produces more agglomerated spherical-shaped particles with irregular distribution throughout the surface. As a result of the higher temperature, irregular particle shapes are fused via adhesion.

Similar to the estimated average crystallite size from the XRD results, the particle is also assumed to be smaller in the sample annealed at 900°C as the particles may be composed of an agglomeration of many crystallites (as marked in Figure 2), which results in the formation of irregular and deformed particle shapes at higher annealing temperatures. Consequently, the smaller the crystallite size, the smaller the particle size. From Figures 2A–C, it seems the grains are dissimilar in different regions. Therefore, the particle size is likely to be different among the samples annealed at 500°C, 700°C, and 900°C as pointed out in Figure 2. For more evidence, the estimated average grain size is about 70 nm for a 500°C annealed sample. Then, it seems the average grain size gets reduced to ∼40 nm at 900°C and is smaller than the sample annealed at 700°C (∼50 nm). This implies that these particles/grains observed in the SEM image are composed of many crystallites/grains. Further, it is believed that the SEM image of the sample annealed at 900°C has composed of small size particles/grains than the other annealed samples.

The elemental composition and valence state of the prepared samples were determined by the XPS technique. Figure 3 shows the recorded XPS data of the sample that was annealed at 900°C. The peak features a, b_, and c are deconvoluted from the recorded C 1s spectra as shown in Figure 3A. The characteristic feature at 284.6 eV is sp2-hybridized carbon (C-C) that was adsorbed on the surface of the prepared sample (Khalid et al., 2013). Additionally, peaks at 286.2 eV, and 288.1 eV are related to the hydroxyl carbon (C-O), and carboxyl carbon (O=C-O), respectively (Liang et al., 2017). The measured O 1s spectra were deconvoluted into three peaks (Features d, e_, and f), as shown in Figure 3B. The presence of various oxygen ions in CuTa₂O₆ with potentially different coordination environments (more Cu-coordinated or more Ta-coordinated), accounts for the deconvolution of the O 1s signals into multiple components. A peak at 529.6 eV (feature d) could be related to the CuO lattice oxygen peak as the binding energy of the lattice oxygen for the bulk samples was frequently found to be slightly higher than that for the nanostructures (Svintsitskiy et al., 2013). Furthermore, the peaks at 530.6 eV and 531.7 eV (features e and f) are associated with Ta−O and Ta=O, respectively. This main peak (feature f) could be attributed to the bulk oxygen ions affected by charge imbalance due to the presence of oxygen vacancy. An additional shoulder peak at 533.0 eV (feature g) is attributed to an oxygen species that is coupled to hydroxyl or hydrated molecules. Consequently, the prepared CuTa₂O₆ phases may contain O-H groups (Chennakesavulu and Ramajaneya Reddy, 2015). The high-resolution pattern of Cu 2p as displayed in Figure 3C exhibits the characteristic peaks at 934.2 eV (a₁ feature) and 954.1 eV (c₁ feature) that correspond to Cu 2p_3/2 and Cu 2p_1/2, respectively, confirming the presence of Cu²⁺. The two other peaks at 943.1 eV (b₁ feature) and 962.3 eV (d₁ feature) are identified as shake-up satellite peaks of Cu 2p_3/2 and Cu 2p_1/2 respectively. The shake-up peaks may have arisen from an unfilled Cu 3d⁹ shell, suggesting the presence of Cu₂O and Cu²⁺ in the prepared sample (Chennakesavulu and Ramajaneya Reddy, 2015) (Muthu Gnana Theresa Nathana et al., 2018). Figure 3D shows two peaks of Ta 4f_7/2 (a₂ feature) and Ta 4f_5/2 (b₂ feature) at 28.2 eV and 29.5 eV indicate the presence of tantalum in the Ta⁵⁺ oxidation state (Wang et al., 2012).

Figure 4A compares the Cu K-edge XANES spectra of the 500°C-, 700°C-, and 900°C-annealed CuTa₂O₆ phase with those of CuO (+1 oxidation state) and Cu₂O (+2 oxidation state) references. The rising-edge peak at ∼8,980 eV for Cu₂O reveals the oxidation state of Cu¹⁺ (Kau et al., 1987). The absence of a pre-edge peak appears in the Cu K-edge XANES spectra is consistent with previous reports (Bera et al., 2002) (Khemthong et al., 2013). The main absorption peak around ∼9,000 eV reveals the 1s-4p transition, which merges into the continuum states (Khemthong et al., 2013). A shoulder peak (feature A) in the energy range 8,985–8,990 eV is attributed to the 1s → 4p_z transition (shakedown). The energy shift of this feature (feature A) from Cu₂O to CuO is about 4.34 eV suggesting the valence state change from Cu¹⁺ to Cu²⁺. The Cu K-edge spectra of CuTa₂O₆ have two regions that include a shoulder peak and a main absorption peak (feature A and feature B). It seems the peak feature (feature A) of all the annealed CuTa₂O₆ is close to that of CuO, revealing the oxidation state of Cu²⁺ (Kau et al., 1987) (Lim et al., 2019). Then, the main absorption peak (feature B) of annealed at 500°C is also similar to those CuO reference samples again confirming the oxidation state of Cu²⁺. However, the main absorption peak (feature B) of annealed 900°C shows a higher intensity compared to the annealed 700 °C suggesting the existence of more oxidation state of Cu²⁺. Also, this main absorption peak of located at a higher energy than the sample that was annealed at 500°C (feature B) so it is believed to have a more oxidation state.

To determine variations in the local atomic environment around Cu in the CuTa₂O₆ catalysts, the Fourier-transformed k³-weighted Cu K-edge spectra were obtained, as shown in Figure 4B. From the 500°C-annealed CuTa₂O₆, the major peaks at 1.56 Å and 3.01 Å correspond to Cu-O and Cu-Cu bonds, respectively (Nasir et al., 2017). The slight contraction in Cu-O bond length and the <comment > shortening in Cu-Cu/Ta bond at </comment>900°C indicates the presence of more oxygen vacancies and cationic vacancies. Commonly, the contracted bond lengths are associated with an increase in the oxidation state. Therefore, the slight contraction in the Cu-O bond at 900°C compared to the 500°C, is also accompanied by the increased oxidation state of Cu. This contraction in the Cu-O bond also suggests the higher coordination number in Cu-O first shell. It is believed that this contraction in Cu-O bond length could be due to the strong interfacial interaction among the Cu-O-Ta materials. It is assumed that the oxidation state of Ta is also to be different in these samples. Particularly, an increase in annealing temperature might lead to an alteration in the valence state of the absorbing atoms in the sample annealed at 900°C. This considerable variation in the coordination of Cu-O and Cu-Cu/Ta suggests the phase transformation of CuTa₂O₆, as evident from the structural analysis. The FT-EXAFS oscillations of the CuTa₂O₆ catalysts exhibit different profiles specifically for sample annealed at 500°C, suggesting dissimilar coordination environments of the Cu-O sites as presented in CuTa₂O₆ in Figure 4C.

The UV-Vis DRS of the prepared CuTa₂O₆ phase was analyzed and is shown in Figure 5. All the prepared samples present a broad absorption range in the visible light region. The redshift that was observed with an increase in annealing temperature suggests that annealing increases sensitivity to visible light. The bandgap energy (E_g) of each prepared catalyst was estimated using the relation E_g = 1,240/λ_g (eV), where λ_g is the absorption edge, which was obtained from the intercept between the tangent of the absorption curve and the abscissa. The estimated band gap energy of the samples that were annealed at 500°C, 700°C, and 900°C corresponds to 3.85 eV, 3.66 eV, and 3.26 eV, respectively. A decrease in the bandgap energy is observed for the sample annealed at a higher temperature (900°C). This could be attributed to the transition of the band from Cu 3d⁺ O 2p orbital to Ta 5d orbital as Cu 3d¹⁰ state may contribute to the valence band of the semiconductor, as similar behavior also reported on the materials such as metal oxides and sulfide photocatalysts (Xu et al., 2012b).

Methyl orange dye was used as a model pollutant. The UV-Vis DRS analysis demonstrated that the prepared CuTa₂O₆ phase was active in the visible region. The photocatalytic process was carried out under visible light irradiation. Figure 6 shows the time-dependent UV-Vis absorption spectra of MO dye solution without and with the addition of H₂O₂ for the samples annealed at different temperatures (a) 500°C, (b) 700°C, and (c) 900°C. The UV-Vis DRS spectra of the MO dye solution show a characteristic peak at 464 nm. From Figure 6; Figure 7, the intensity of this characteristic peak gets decreases as the duration of exposure to light increases. At 90 min of exposure, the intensity of the peak was significantly reduced in comparison with 0 min, and the aqueous solution was colorless, indicating the decomposition of the dye molecules. The continuous disappearance of the absorption band suggests that the functional group of the MO dye is broken down (Xu et al., 2012b).

Figure 8A displays the estimated degradation efficiency of MO dye in the absence of H₂O₂. The percentage degradation efficiency was calculated as, D e g r a d a t i o n % = C 0 − C C 0 × 100 % where C₀ is the initial concentration of MO dye and C is the concentration of the dye following irradiation for time t. The graph in Figure 8A displays the degradation efficiencies of the samples annealed at 500°C, 700°C, and 900°C were found to 70.4%, 74.37%, and 81.3%, respectively. The sample annealed at 900°C shows the highest degradation efficiency revealing its superior photocatalytic activity. The increase in photodegradation efficiency can be attributed to a decrease in crystallite size, as agrees with the XRD results. Crystallite shrinkage increases the surface area-to-volume ratio of the catalyst, increasing the number of reactive sites and surface hydroxyl groups (Li et al., 2011). A large surface-area-to-volume ratio favors the reaction/interaction between the photocatalyst and the dye molecules, resulting in higher degradation efficiency. XRD results evidences that the samples annealed at 700°C and 900°C have similar phase structures and crystallinities. However, the sample annealed at 900°C exhibit a better photocatalytic performance whoci could be due to the existence of smaller crystallite size and more surface defects in materials. Smaller crystallite size corresponds to a larger total number of atoms with unsaturated coordinates on the surface, and these atoms have significantly improved photocatalytic activity. Additionally, the recombination of electrons and holes decreases as the crystallite size reduces, and thus the generated electrons and holes are transferred readily to the surface of the catalyst, favoring the process (Mageshwari et al., 2013).

To examine the surface area and porous structure of the prepared CuTa₂O₆ samples, BET surface analysis was carried out using nitrogen isothermal adsorption-desorption. Supplementary Figures S2A–C displays the isotherm curves of CuTa₂O₆ at different annealed temperatures along with the estimated pore-size distribution using the BJH method (Supplementary Figures S2D–F). The calculated specific surface area and BJH desorption Pore volume are presented in Supplementary Table S1. Generally, the surface area to volume ratio of nanomaterial plays a substantial role in influencing the material’s properties (Xia et al., 2012). When the particle size decreases, there is an enhancement in the surface area-to-volume ratio (Liu et al., 2015). Consequently, the sample annealed at 900°C shows a higher BET surface area (10.96 m²/g) than the other samples (9.07 m²/g for 500°C and 7.56 m²/g for 700°C), indicating a greater number of existing active sites. We, therefore, believe that the large specific surface area benefits better access and diffusion of liquid and gaseous reactants, which is beneficial for photocatalytic reactivity. The increased pore size distribution in the BJH desorption process is also likely to improve the photocatalytic performance of the materials (Supplementary Figures S2D–F).

The kinetics of the process importantly affect the rate at which pollution is removed. The kinetics of photocatalysis were determined from experimental data using the Langmuir—Hinshelwood model. ln C 0 C = K r t

Here, K_r is the degradation rate constant and t is the degradation time. The degradation rate constant K_r is calculated from the slope of the kinetic plot of the natural logarithm of the concentration ratio ln (C/C₀) versus the irradiation time t in minutes. Figure 8B plots this curve for MO dye in the absence of H₂O₂. The concentration ratio varies linearly with time suggest that the photodegradation of MO dye follows first-order kinetics. The degradation rate constants K_rt for the samples that were annealed at 500°C, 700°C, and 900°C are calculated to be 0.01751 min^-1, 0.01333 min^-1, and 0.01107 min^-1, respectively.

To enhance the photocatalytic activity of the prepared CuTa₂O₆, H₂O₂ (0.6 mL) was added as a green additive. Figure 8C shows the degradation efficiency of the samples with the addition of H₂O₂; they have higher degradation efficiencies than those achieved without H₂O_2. Therefore, the addition of H₂O₂ substantially enhanced photocatalytic activity. Then, Figure 8D displays the linear increase in the degradation rate. The degradation efficiencies of the samples that were annealed at 500°C, 700°C, and 900°C with added H₂O₂ are calculated to be 82.5%, 87.3%, and 94.7% with estimated rates constants of 0.01251 min^-1, 0.01129 min^-1 and 0.00866 min^-1, respectively. Hence, the addition of H₂O₂ achieves the efficient photodegradation of MO in the presence of copper tantalum oxide.

The high-performance activity as a result of adding H₂O₂ can be explained as follows. H₂O₂ acts as a good electron acceptor, accelerating the reaction by increasing the formation of oxidizing radicals, which rapidly degrade the dye molecules. When the photons are incident upon the semiconductor, the photogenerated holes react with H₂O or OH⁻ and produce hydroxyl radicals (^•OH) while the superoxide (O₂ ^−•) and (^•OH) radicals are generated by the electrons in the conduction band (e_CB ⁻). The produced O₂ ^−• radicals further form hydroperoxy (OOH^•) and ^•OH radicals, which destroy the organic contaminants. The possible chemical reactions for the degradation of organic dyes are as follows. S e m i c o n d u c t o r + h γ   →   h V B + + e C B − h V B + + H 2 O   → O − H + H + h V B + + O H −   → O • H e C B − + H 2 O 2   → O • H + O − H e C B − + O 2   →   O 2 − • O 2 − • + H +   →   H O 2 • H 2 O 2 + O 2 − •   → O • H + O H − + O 2 H 2 O 2 + h γ   →   2 H 2 O D y e + ( • O H / O 2 − • / O O H • )   →   d e g r a d a t i o n

To investigate the stability and reusability of the catalysts, an experiment on the recycling of the samples that were annealed at 500°C and 900°C was conducted. The catalyst sample annealed at 900°C underwent little change or deactivation, as shown in Figure 9A. The sample annealed at 500°C exhibits a considerable decrease in the efficiency of photodegradation, as shown in Figure 9B. Notably, the photodegradation efficiency experiences a rapid decline as early as the second and third runs. Therefore, it is believed that the cubic phase of CuTa₂O₆ is more stable for the photodegradation of MO dye than the orthorhombic phase (sample that was annealed at 500°C).