Artificial photosynthesis to directly convert light energy into chemical energy currently attracts much interest in both fundamental study and applications ^{[1, 2, 3, 4, 5]}. The terminal step in Photosystem II (PSII) and a possible half-reaction for artificial photosynthesis is water oxidation,

which is a challenging reaction ^{[6, 7, 8]} that is a significant mechanistic challenge in its requirements for extracting both electrons and protons to reduce protons or other chemicals to produce fuel and O-O bond formation. Much effort has been devoted to identify new catalysts made of cheap and abundant materials for water oxidation, to integrate them into photosynthetic schemes. Water oxidation catalysts that contain common metals such as Co ^{[9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]}, Mn ^{[25, 26, 27, 28]}, Fe ^{[29, 30, 31, 32]}, Cu ^{[33, 34, 35, 36, 37, 38]}, V ^[39], and Ni ^{[40, 41, 42, 43, 44, 45, 46]} have been studied under photo- and electrochemical conditions. Among the first row transition metal (TM) elements, Cu is the eighth most abundant element in the Earth’s crust and Cu-based materials have been suggested for water oxidation catalysts (WOCs) for both their well-defined coordination chemistry and useful redox chemistry. However, little success has been achieved in the search for an efficient Cu-based WOCs. No CuO has been reported as a photocatalytic water oxidation catalyst yet.

Copper complexes in the higher oxidation state are well-known catalysts for oxidizing organic compounds such as alcohols, phenols and hydrocarbon. The use of Cu for water oxidation has been much less explored. Recently, Mayer et al. ^[33] reported the first example of a homogeneous copper electrocatalyst for water oxidation. Meyer et al. ^[47] reported that simple Cu(II) salts were highly robust and sustainable electrocatalytic water oxidation catalysts. Lin et al. ^[37] produced a highly active electrocatalytic water oxidation catalyst by simply mixing a Cu(II) salt and 6,6-dihydroxy-2,2-bipyridine (H₂L) in a basic aqueous solution. Sun et al. ^[36] prepared a robust electrocatalytic water oxidation catalyst based on copper oxide by facile electrodeposition of Cu²⁺ from borate buffer solution under near neutral conditions. However, it is still a continuing challenge to design efficient, easily accessible, and stable catalysts for the photocatalytic water oxidation reaction that is based on the common copper metal.

Here, flower-like three dimensional (3D) CuO microspheres were easily prepared from a Cu(NO₃)₂·3H₂O precursor. We report on the photoinduced water oxidation obtained by using a sacrificial system made of [Ru(bpy)₃]²⁺ (bpy = 2,2′-bipyridine) as a photosensitizer, Na₂S₂O₈ as the sacrificial acceptor, and the flower-like 3D CuO microspheres as the photocatalytic water oxidation catalyst. The catalytic activity of the CuO microspheres in borate buffer showed a best performance of a TOF of 0.21 μmol s^-1 m^-2. Our data clearly demonstrated that the evolved O originated exclusively from the water. We suggest that the catalytic activity for O₂ evolution can be related to the enhanced specific surface area of CuO.

H₂¹⁸O (97% ¹⁸O) was purchased from Mashall Isotope Ltd. Purified water (18.2 MΩ·cm) for the preparation of the solutions was obtained from Molecular Lab Water Purifier, and all other chemicals and salts used were of the highest purity available from commercial sources. All chemicals were of analytical grade and used without further purification. In a typical synthesis, 1 g of block polymer P123 was dissolved in 40 mL distilled water. Cu(NO₃)₂·3H₂O (0.005 mol) was added to the solution and this mixture was stirred until a clear blue solution was formed. Then, ammonia solution (10 mL, 25-28 wt%) was added, which was followed by the formation of a dark blue solution. The resulting dark blue mixture was stirred for 1 h,transferred into a 100 mL three-necked flask, and refluxed at 110 °C for 2 h. Vigorous stirring was maintained throughout the process. After the reaction, the system was cooled naturally to room temperature. Precipitates were collected by centrifugation and washed several times with deionized water and ethanol. The resulting precipitates were dried at 80 °C (yield 0.2 g, 41%).

Photocatalytic water oxidation was performed as follows. The desired concentration of catalyst 0.5 g/L was added to a buffer solution (80 mmol/L, pH 8.5 for borate buffer) containing [Ru(bpy)₃]Cl₂ (1.0 mmol/L) and Na₂S₂O₈ (5.0 mmol/L). The above solution was deaerated by purging with Ar gas for 10 min in a flask (28 mL) sealed with a rubber septum (the volume of the reaction solution was 15 mL). The reaction was then started by irradiating the solution with a LED light source (light intensity 16 mW, beam diameter 2 cm) (Fig. S1) through a transmitting glass filter (λ ≥ 420 nm) at room temperature. After each irradiation time, 150 μL of Ar was injected into the flask and then the same volume of gas in the headspace of the flask was withdrawn by a SGE gas-tight syringe and analysed by gas chromatography (GC). The O₂ in the sampled gas was separated by passing through a packed molecular sieve 5A column (2 m × 3 mm) with an Ar carrier gas and quantified by a thermal conductivity detector (TCD, Shimadzu GC-9A). The total amount of evolved O₂ was calculated from the concentration of O₂ in the headspace gas. Contamination of the head-space with air was corrected by measuring of N₂ present in the head-space (from the N₂ peak in the GC traces). The solution pH was measured after the reaction by a METTLER TOLEDO FEP20 pH meter.

X-ray photoelectron spectra (XPS) were measured by ESCALAB250xi with X-Ray monochromatisation. GC-MS spectral analyses of isotopic labelled O₂ were performed on an Agilent Series Model 7890A chromatograph interfaced with an Agilent Series Model 5975C mass spectrometer. X-ray diffraction (XRD) data were collected with a PANalytical X'Pert Pro Diffractometer operated at 40 kV and 40 mA with Cu K_α radiation (step size: 0.017°, step time: 10.34 s). Scanning electron microscopy (SEM) observations were performed on a Hitachi S-4800 microscope operated at an accelerating voltage of 5.0 kV. Transmission electron microscopy (TEM) images were obtained with a JEOL JEM-2010 instrument operated at 200 kV. N₂ adsorption at -196 °C was performed with a Micromeritics ASAP 2020M system. A sample mass of about 100 mg was used for the adsorption analysis after pretreatment at 80 °C for 8.0 h under vacuum conditions and kept in N₂ atmosphere until the N₂ adsorption measurement. The surface area was calculated using Langmuir and Brunauer-Emmett-Teller (BET) method.

Cyclic voltammetry (CV) was recorded on a CHI660D electrochemical analyser. A glassy carbon, an Ag/AgCl and a Pt wire electrodes were used as a working, reference and auxiliary electrodes, respectively. CV was obtained in buffer solutions at room temperature with a scanning rate of 100 mV /s.

The 10.8 atom % H₂¹⁸O of borate buffer solution (pH 8.5, 80 mmol/L) containing CuO (0.5 g/L), [Ru(bpy)₃]Cl₂ (1.0 mmol/L), Na₂S₂O₈ (5.0 mmol/L) was deaerated with Helium gas before irradiation by LED light (λ ≥ 420 nm) in a flask that is sealed with a rubber septum. After 9 min, 50 μL of gas sample was withdrawn using a gas-tight syringe for gas analysis. An Agilent Series 7890A model chromatograph interfaced with an Agilent Series 5975C model mass spectrometer operating in electron impact ionization mode was used to collect mass spectrometric data. The MS detector was tuned for maximum sensitivity (quadruple temperature, 150 °C; ion source temperature, 230 °C). The single ion mode was used to scan for the ions m/z = 28, 32, 34, 36 with a dwell time of 100 ms, resulting in 8.3 cycles per second. The ions of m/z range from 30 to 50 were also scanned in order to observe the abundance change of ¹⁶O¹⁸O and ¹⁸O¹⁸O, which evolved from H₂¹⁶O and H₂¹⁸O, respectively. The total flow rate into the spectrometer was limited to 0.6 mL/min. The GC equipped with a molecular sieve column (30 m × 0.32 mm × 15 μm), and the vaporizing chamber temperature and column temperature was set for 100 and 35 °C, respectively.

CuO microspheres were synthesized by a modified literature method ^[48]. The morphology of the synthesized CuO microspheres was examined by SEM and TEM. Fig. 1(a) shows typical SEM image of the synthesized CuO microspheres. It showed a quite uniform microsphere structure with an average diameter of about 2.0 µm. In Fig. 1(b), the morphology of the CuO microsphere material is seen to be self-assembled radially comprising many nanoplates with a thickness of about 25 nm. Fig. 1(c) shows a typical TEM image of the CuO spheres. To confirm the formation of CuO, the material was characterized by powder XRD (Fig. 1(d)). All the peaks of the CuO microspheres can be assigned to monoclinic symmetry of CuO (JCPDS 05-0661). The main peaks were located at 2θ = 32.6°, 35.6°, 38.7°, 48.7°, 53.4°, 58.3°, 61.6°, 66.2°, 68.2°,72.4°and 75.1°, which correspond, respectively, to the (110), (−111), (111), (−202), (020), (202), (−113), (−311), (220), (311) and (−222) planes of monoclinic CuO. No impurity peak was observed, indicating the high purity of the prepared samples. The broad XRD peaks also indicated the sample was composed of CuO nanocrystallines. The electron dispersive spectroscopy (EDS, Fig. S2) spectrum further confirmed the existence of Cu and O in the composites. The crystal structure and coordination mode of monoclinic CuO are shown in Fig. 2(a) and (b). The Cu−O distance was 1.95 and 1.96 Å. FT-IR spectrum also confirmed that the prepared CuO samples are a pure phase of CuO with the monoclinic structure (Fig. S3). Raman spectrum also confirmed that CuO samples are a pure phase (Fig. S4). The composition and valence states of the CuO were further investigated by XPS (Fig. S5). XPS (Fig. S5(a)) detected the Cu 2p_3/2 peaks at 933.7 eV with two satellite peaks at 941.0 and 943.5 eV and Cu 2p_1/2 peaks at 953.7 eV with a satellite peak at 962.5 eV, respectively. These features correspond to a Cu²⁺ state for the Cu atoms. As shown in Fig. S5(b), the O 1s core level spectrum was broad, and two peaks (marked as I and II) were observed. Peak I at the lower energy of 529.6 eV is in agreement with that for O^2- in CuO, while peak II at the higher energy of 531.4 eV was attributed to oxygen adsorbed on the surface of the CuO microspheres. The XPS results further support the conclusion that the sample is pure CuO.

Fig. 1. SEM images (a, b), TEM image (c), and XRD pattern (d) of CuO microspheres.

Fig. 2. Crystal structure of monoclinic CuO (a) and its coordination mode (b).

The cyclic voltammogram of the CuO microspheres in a pH 5.8 buffer solution showed a catalytic water oxidation peak (the reference current was zero) at 1.82 V vs. RHE (Fig. 3(a)). This potential was much higher than the potential of the oxidized state of [Ru(bpy)₃]^2+/3+ (E_1/2 = 1.67 V vs. RHE). Consequently, light driven water oxidation by CuO microspheres using [Ru(bpy)₃]²⁺ as a photosensitizer was thermodynamically unfavored in a pH 5.8 buffer solution. Water oxidation is favoured under basic conditions. Therefore, the CuO microspheres were potentially useful for water oxidation under near neutral conditions. The electrochemical properties of the CuO microspheres were investigated in pH 8.5 borate buffer solutions (Fig. 3(b)). A catalytic water oxidation curve began from 1.72 V vs. RHE, which were 0.10 V vs. RHE lower than that observed in a pH 5.8 buffer solutions. In contrast, only the minimum current was obtained in the same buffered solution without the CuO microspheres (Fig. 3(b)). Most importantly, the catalytic potential of the CuO microspheres was lower than that of [Ru(bpy)₃]^2+/3+, which indicated that [Ru(bpy)₃]^2+/3+ could drive the CuO microspheres to oxidize water.

Fig. 3. (a) CVs of 80 mmol/L sodium Na2SiF6 + NaHCO3 solution at pH 5.8 using 0.5 g/L CuO microspheres and 1.0 mmol/L [Ru(bpy)3]Cl2. (b) CVs of 80 mmol/L sodium borate buffer solution at pH 8.5 with 0.5 g/L of CuO microspheres and 1.0 mmol/L of [Ru(bpy)3]Cl2.

Light-driven water oxidation was performed by the photoirradiation (λ ≥ 420 nm) of a borate buffer (80 mmol/L, pH 8.0-9.5, 15 mL) containing the metal oxide catalyst, Na₂S₂O₈ as the sacrificial electron acceptor, and [Ru(bpy)₃]Cl₂ as the photosensitizer. The catalytic cycle of this visible light-driven water oxidation is depicted in Scheme S1. Isotope-labelled water oxidation experiments using ¹⁸O-enriched water (10.8%) instead of H₂¹⁶O were carried out with the CuO microsphere catalyst to confirm that the water was the source of the evolved oxygen. The relative abundance of oxygen isotopes, which were determined from the intensities of their three molecular ion peaks, is listed on Fig. S6. The ratio of ¹⁶O¹⁶O:¹⁶O¹⁸O:¹⁸O¹⁸O was determined to be 81:18:1, which is in good agreement with the simulated ratio of 80:19:1 when the oxygen comes exclusively from the water. The data clearly demonstrated that the evolved O originated exclusively from the water.

Water oxidation is pH dependent and the catalytic activity of the CuO microspheres was studied under different pH conditions (Fig. 4). The reaction at pH 8.5 showed the highest O₂ yield and turnover frequency (TOF). The O₂ yield and TOF decreased with increasing pH from 8.5 to 9.5. Although a high pH is thermodynamically favourable for water oxidation, a higher pH can also accelerate the degradation of the photosensitizer, which is a competitive process in the reduction of [Ru(bpy)₃]³⁺ to [Ru(bpy)₃]^{2+ [10, 49, 50]}. When pH 9.0 and 9.5 were used, the orange red colour of the fresh reaction solution turned much darker after O₂ evolution stopped. The higher pH value promoted the degradation of the photosensitizer by the nucleophilic attack of water or OH^- on [Ru(bpy)₃]³⁺ before the photocatalytic water oxidation. This was supported by UV-vis evidence (Figs. S7 and S8). When the pH 8.0 buffer was used, the O₂ evolution and TOF also decreased relative to that of the pH 8.5 buffer. The orange red color of the fresh reaction solution turned green (Fig. S9) after 30 s of photoirradiation. The solution colour turned green with reaction time, demonstrating that there was a pile up of [Ru(bpy)₃]³⁺ and Eq. (2) became the rate determining step.

Fig. 4. Kinetics of O2 formation in the photocatalytic system under various pH conditions using CuO microspheres (1) pH 5.8, Na2SiF6+NaHCO3, 20 mmol/L; (2) pH 8.0, 80 mmol/L NaBi; (3) pH 8.5, 80 mmol/L NaBi; (4) pH 9.0, 80 mmol/L NaBi; (5) pH 9.5, 80 mmol/L NaBi. Reaction conditions: LED lamp (λ ≥ 420 nm), catalyst concentration 0.5 g/L, [Ru(bpy)3]Cl2 1.0 mmol/L, Na2S2O8 5.0 mmol/L, total reaction volume 15 mL, overall volume ~28 mL, with vigorous agitation using a magnetic stirrer.

The presence of [Ru(bpy)₃]³⁺ was proved by UV-vis by its characteristic absorbance at 670 nm (Fig. S10). First, the existence of accumulated [Ru(bpy)₃]³⁺ meant that a large amount of the water oxidation oxidant could not be effectually used in the catalytic cycle. Second, at pH 8.0, the pH value of the reaction solution was measured to decrease to 2.7 after the reaction stopped as a result of the release of protons. This demonstrated that the acetic-alkali equilibrium was insufficient for maintaining a thermodynamically favourable pH value for photocatalytic water oxidation. Moreover, when pH 8.0 was used, the time of O₂ evolution stopped earlier. It achieved a plateau within 3 min, which demonstrated that it was thermodynamically unfavorable in a pH 2.7 buffer solution. In other words, the light-driven water oxidation by the CuO microspheres using [Ru(bpy)₃]²⁺ as a photosensitizer was thermodynamically more favourable for photocatalytic water oxidation at higher pH values. These results indicated that O₂ evolution depended on a combination of several factors, which include low pH, electron acceptor consumption, photosensitizer (Fig. S11) and catalyst deactivation.

The photocatalytic water oxidation was investigated with different concentrations of the CuO microspheres in 28 mL total volume. The catalytic performances are compared in Fig. 5. O₂ rapidly formed after 1 min of visible light illumination, and the O₂ evolution rate decreased with time (the slope of the curve became level with the lapse of time), and O₂ evolution achieved a plateau in 11 min. The reason is that the sacrificial electron acceptor of Na₂S₂O₈ was partially used up and the photosensitizer of [Ru(bpy)₃]Cl₂ was decomposed during the photocatalytic process, so the amount of O₂ formation was limited. A maximum O₂ evolution yield (yield was defined as twice the number of moles of O₂ per mole of Na₂S₂O₈) of 11.5% and O₂ evolution amount of 4.4 μmol were obtained when the concentration of the CuO microspheres was 0.5 g/L. The O₂ yield increased with the catalyst concentration up to 1.0 g/L. The yield for O₂ evolution increased to 12.0%, and decreased with a lower catalyst concentration of 0.25 g/L. The yield for O₂ evolution decreased to 5.7%. Moreover, the TOF decreased from 2.5× 10^-4 to 2.1 × 10^-4 and 0.9 × 10^-4 s^-1 per Cu. The amount of O₂ increased with catalyst concentration, but it reached saturation at 1.0 g/L.

Fig. 5. Kinetics of O2 formation in the photocatalytic system using different concentrations of CuO microspheres. The other conditions were the same as in Fig. 4 (pH 8.5, 80 mmol/L NaBi).

The visible light-driven water oxidation catalyzed by the CuO microspheres was investigated in different buffers: borate buffer (pH 8.5, 80 mmol/L) and phosphate buffer (pH 8.5, 80 mmol/L, Figs. S12 and S13). The catalytic activity of the CuO microspheres in borate buffer showed the highest catalytic activity with an O₂ yield of 11.5%. The O₂ evolution yield decreased to 4.8% in the phosphate buffer (H₂PO₄^- and HPO₄^2-). When the phosphate buffer was used, the pH value of the reaction solution changed a lot after the oxygen evolution was completed. The final pH value was decreased to 6.5 with the phosphate buffer, whereas the final pH was decreased to 8.2 for the borate buffer. A lower amount of O₂ evolution was observed for the reaction in the phosphate buffers. We think that this can be ascribed to its low acetic-alkali equilibrium because water oxidation is thermodynamically unfavourable under low pH conditions. In addition, phosphate accelerates the decomposition of [Ru(bpy)₃]³⁺ (Fig. S14).

To confirm that the oxygen evolved in the photocatalytic water oxidation reaction was indeed catalyzed by the CuO microspheres, several control experiments (Table S1) were used. (1) Under identical conditions without the photosensitizer, no oxygen was evolved in the presence of light (Fig. S15(2)), confirming that [Ru(bpy)₃]²⁺ is an essential ingredient for water oxidation in the photocatalytic system. (2) Without the CuO microspheres catalyst, little oxygen was generated in the presence of Na₂S₂O₈ and [Ru(bpy)₃]Cl₂ in a borate buffer (pH 8.5) under photo-irradiation (Fig. S15(3)). The oxygen evolution amount is one-ninth that in the catalytic runs using 0.5 g/L CuO microspheres, indicating that the CuO microspheres were indeed involved in the catalytic oxygen evolution processes. (3) Without the sacrificial electron acceptor, no oxygen was generated in the presence of light (Fig. S15(1)), revealing that Na₂S₂O₈ is an essential ingredient for water oxidation in the photocatalytic system. (4) An irradiation control experiment showed that the catalytic oxygen evolution in the system was driven by light (Fig. S16). All these results clearly demonstrated that the light-driven water oxidation was catalyzed by the CuO microspheres.

The time course of O₂ evolution with different metal oxides is shown in Fig. 6. Table 1 displays the yield of O₂ obtained from the metal oxides studied. A minimum amount of O₂ evolution was detected in the absence of a catalyst. The yield of O₂ evolution obtained after 9 min of photoirradiation with the CuO microspheres (11.5%) was higher than those with Cu(NO₃)₂^.3H₂O (10.7%) and commercial CuO (2.1%). The BET surface area (A_BET) of the CuO microspheres (15 m²/g) was significantly higher (5 times) than that of the commercial CuO (3 m²/g). Clearly, the CuO microspheres have a superior catalytic activity to that of the commercial CuO under photocatalytic water oxidation. Jiao et al. ^[22] reported a novel method to synthesize free unsupported ultra-small cobalt oxide nanometres materials by an in situ phase transformation approach with a layered Co(OH)(OCH₃) precursor. The cobalt oxide with a particle size less than 2 nm gave a TOF of 0.023 s^-1 per cobalt in photocatalytic water oxidation. However, 6 nm Co₃O₄ nanoparticles gave a lower TOF of 0.002 s^-1 per cobalt in photocatalytic water oxidation. X-ray absorption measurements indicated a distinct nanocubane structure, where 13 Co atoms were fully coordinated with O atom by a six-coordinate pattern to form 8 Co₄O₄ cubanes, which may explain the unusually robust water oxidation catalytic activity. So, a smaller particle size and Co³⁺ produce better catalytic activity.

Fig. 6. Kinetics of O2 formation in the photocatalytic system using different catalysts. Other conditions were the same as in Fig. 4 (pH 8.5, 80 mmol/L NaBi).

Table 1 Water oxidation catalyzed by different catalysts. Catalyst ABET (m2/g) O2 yield (%) TOF a (μmol s-1 m-2) TOF b (s-1 per Cu) CuO microspheres 15 11.5 0.21 2.5 × 10-4 Cu(NO3)2·3H2O c — 10.7 — — Commercial CuO 3 2.1 0.17 0.4 × 10-4 No catalyst — 1.4 0 0 Reaction conditions: LED lamp (λ ≥ 420 nm); catalyst concentration 0.5 g/L, [Ru(bpy)3]Cl2 1.0 mmol/L, Na2S2O8 5.0 mmol/L, sodium borate buffer 80 mmol/L (initial pH 8.5); total reaction volume 15 mL, overall volume ~28 mL, with vigorous agitation using a magnetic stirrer. a Normalized by catalyst surface area for O2 evolution in one minute. b Based on the oxygen product rate (in 1 min) divided by the Cu atoms in the catalyst. c 2.7mmol/L (equivalent to 0.5 g/L).

Table 1 Water oxidation catalyzed by different catalysts.

Here, the valence of Cu in the CuO microspheres and commercial CuO was +2 by the XPS analysis. The CuO microspheres (A_BET = 15 m²/g) gave a superior catalytic activity (TOF = 2.5×10^-4s^-1 per Cu) to that of commercial CuO (A_BET = 3 m²/g, TOF = 0.4×10^-4 s^-1 per Cu) for photochemical water oxidation. The results suggested that the photocatalytic activity of CuO depended mainly on the surface area of the catalyst.

A fresh buffer solution containing [Ru(bpy)₃]Cl₂ (1.0 mmol/L) and Na₂S₂O₈ (5.0 mmol/L) was added to the recovered particles for the repetitive examination under photoirradiation. The high catalytic activity of CuO microspheres was maintained even after the fifth run (Fig. 7). To further check the stability of the CuO catalyst, XRD (Fig. S17), FT-IR (Fig. S18), TEM (Fig. S19) and Raman (Fig. S20) measurements were performed after the photocatalytic water oxidation. Fig. S17 shows the diffraction peaks of the CuO after reaction. The typical CuO diffraction peaks remained and were the same as those before the photocatalytic water oxidation. In addition, FT-IR spectra were used to investigate the surface properties of the CuO. As shown in Fig. S18, there are three peaks observed at ~450, 512 and 592 cm^-1, which are the characteristic stretching vibrations of Cu-O bond in monoclinic CuO. None of the peaks showed a change after the photocatalytic water oxidation. No variation in the morphology of the CuO microsphere catalyst was observed (Fig. S19) before and after the reaction. In addition, a catalytic test of the supernatant solution was performed. No O₂ was detected after the catalytic tests by the addition of [Ru(bpy)₃]Cl₂ and Na₂S₂O₈ to the supernatant solution, which excluded Cu²⁺ leaching as a source of smaller secondary (amorphous) nano-catalysts. However,particle was detected when Cu(NO₃)₂· 3H₂O was used as photocatalytic water oxidation catalyst (Fig. S21). These results clearly indicated that the CuO microspheres a highly active and robust catalyst in the photocatalytic system. Our earlier work ^[38] revealed that the surface of the catalyst after the photochemical measurements could show micro-phase separation and deactivation of the catalyst. A series of characterization were performed based on this knowledge. We carried out XPS measurements after the reaction to understand this surface phenomenon. XPS was performed to check the valence states and surface properties of the CuO sample after the photocatalytic reaction. Fig. 8(a) and (b) shows the high resolution XPS spectra of Cu 2p and O1s before and after reaction. The binding energy of each element was calibrated by the C 1s peak (284.8 eV). Fig. 8(a) shows the XPS spectra for Cu 2p_3/2 peaks at 933.7 eV with two satellite peaks at 941.0 and 943.5 eV and Cu 2p_1/2 peaks at 953.7 eV with a satellite peak at 962.5 eV for the CuO microspheres samples before and after the reaction. The binding energies of these peaks indicated the Cu species in the sample was Cu²⁺ by comparison with the peak positions of the Cu 2p peaks of CuO ^[51]. Fig. 8(b) exhibits that the XPS spectrum had the O 1s peak at 529.6 eV with another peak at 531.4 eV for the CuO microspheres before the reaction and the peaks were at 529.7 eV with another peak at 531.4 eV after the reaction. Although the main O 1s peak from the sample after the reaction was slightly shifted to higher binding energy, the same separation between the main peak and satellite peak in both samples and the similarity of peak shapes including the satellite peaks in the whole energy region between 528 and 536 eV strongly suggested that there was no change in the valence state of O^2-. Therefore, the XRD, FT-IR, and XPS results indicated that the CuO material was a highly robust catalyst in the photocatalytic water oxidation.

Fig. 7. Kinetics of O2 formation in the photocatalytic system using fresh CuO microspheres and recovered CuO microspheres. Other conditions were the same as in Fig. 4 (pH 8.5, 80 mmol/L NaBi).

Fig. 8. XPS of CuO microspheres before and after the reaction in the energy regions of Cu 2p (a) and O 1s (b) of CuO microspheres before and after the reaction.

In a recent mechanistic study, water oxidation intermediates were observed using a time-resolved FT-IR technique. Frei and co-workers ^[52] studied the behaviour of Co₃O₄ nano-particles in the visible light driven [Ru(bpy)₃]²⁺-S₂O₈²⁻ system and proposed two distinct active sites, namely oxo-bridged Co(III)OH binuclear site and single Co(III)OH sites. The time-resolved FT-IR measurements demonstrated that the presence of adjacent CoOH groups accompanied an oxygen bridge led to an active site with a TOF higher than 3 s⁻¹, while the single cobalt site exhibited a much slower TOF (∼0.02 s⁻¹ per surface atom). Cu^II has both suitable redox properties and coordination pattern based on the oxidation to Cu^III or even Cu^IV and reduction to Cu⁰ or Cu^{I [53, 54, 55, 56]}. With a propensity for square planar coordination, d⁸ Cu^III is active in reactions of copper compounds containing organic ligands and in bis(m-oxo)-bridged complexes. Cu^IV complexes stabilized either by fluoride ligands or as linear O=Cu=O are also known. Copper complexes of a high oxidation state have been shown to oxidize phenols ^[54], alcohols ^[55] and even hydrocarbons ^[53]. In our CuO microsphere photocatalytic water oxidation system, a TOF of 2.5 × 10^-4(s^-1 per Cu) was observed under optimal condition. This can be related to the special property of CuO because it can undergo a change of the oxidation state in the oxygen evolution reaction (OER) that is similar to cobalt oxide ^[57]. Therefore, we suggest that regardless of the presence of Cu-O-Cu, the dominant reaction pathway for the copper oxide catalyst is the single Cu(II)OH site mechanism.

There are several reasons why a significantly high O₂ yield has not been obtained yet. For example, electron injection from an adjacent Cu(II)OH site to the [Ru(bpy)₃]²⁺ sensitizer can be difficult due to the short Cu−O distance (1.95 or 1.96 Å) compared to the [Ru(bpy)₃]²⁺ complex (11 Å). In addition, electron transfer from the Cu(II)OH active site to the [Ru(bpy)₃]²⁺ sensitizer is a random process with no control of the direction, which also reduce the synergism of two adjacent Cu atoms. A precise control of electron transfer has to be optimized to achieve a high O₂ yield. Third, the catalytic potential of the CuO microspheres was only 0.13 V lower than that of [Ru(bpy)₃]²⁺, which indicated that [Ru(bpy)₃]³⁺ only weakly drive the CuO microspheres to oxidize water. Further investigation will be carrying out to fully determine the water oxidation reaction mechanism on the surface of CuO microspheres.

This is the first report where a copper species was demonstrated to be a photocatalytic water oxidation catalyst under near neutral conditions. The CuO catalyst exhibited excellent water oxidation activity and stability. No change was found by XPS analysis in the surface condition of the CuO microspheres before and after the photocatalytic reaction. Copper is an attractive material because of its abundance and lower price than other elements such as Co, Ni, and Mo. A photocatalytic water oxidation reaction mechanism catalyzed by the CuO microspheres was proposed. Easy preparation, low toxicity, and rich redox properties render the Cu-based photocatalyst favorable for future water splitting applications.

Table S1 Water oxidation catalyzed without CuO or Ru(bpy)3Cl2 or persulfate or light. Entry [Ru(bpy)3]Cl2 (mmol/L) Na2S2O8 (mmol/L) Catalyst (g/L) light O2 (μmol) 1 1 5 0 yes 0.5 2 1 0 0.5 yes 0 3 0 5 0.5 yes 0 4 1 5 0.5 no 0 Conditions: LED lamp (λ ≥ 420 nm), sodium borate buffer (initial pH 8.5) 80 mmol/L, total reaction volume 15 mL, overall volume ~28 mL, vigorous agitation using a magnetic stirrer.

Table S1 Water oxidation catalyzed without CuO or Ru(bpy)3Cl2 or persulfate or light.

Scheme S1. Spectrum of the LED lamp.

Fig. S1. Spectrum of the LED lamp.

Fig. S2. EDS of the CuO microspheres.

Fig. S3. FT-IR spectrum of the products of the CuO microspheres. There were three infrared peaks observed at ~450, 512 and 592 cm-1, which are the characteristic stretching vibrations of Cu-O bond in monoclinic CuO. The high-frequency mode at ~592 cm-1 may be a Cu-O stretching along the [-101] direction, and the mode ate ~512 cm-1 may be Cu-O stretching along the [101]. These data are same as reported data.

Fig. S4. Raman spectrum of CuO microspheres excited by 514.5 nm laser.

Fig. 14. Fig. S5. XPS spectra of CuO microspheres. (a) Cu 2p; (b) O 1s.

Fig. S6. Observed and theoretical relative abundances of 18O-labeled and unlabeled oxygen evolved during the photocatalytic oxidation of a buffer solution (4.5 mL) prepared with H218O-enriched water (10.8% H218O) containing CuO (0.5 g/L), [Ru(bpy)3]Cl2 (1.0 mmol/L) and Na2S2O8 (5.0 mmol/L) Exp.—observed mass intensity; Cal. —calculated values assuming that evolved O2 results exclusively from water.

Fig. S7. UV-vis spectral changes during the photocatalytic O2 evolution with CuO microspheres. The black line shows the absorption of aqueous borate buffer solutions (pH 8.5, 80 mmol/L) containing [Ru(bpy)3]Cl2 (1.0 mmol/L), Na2S2O8 (5.0 mmol/L) and CuO microspheres (0.5 g/L). The red line shows the absorption of above solution after 9 min of irradiation. The concentration of [Ru(bpy)3]2+ decreased by 6.4%.

Fig. S8. UV-vis spectral changes during the photocatalytic O2 evolution with CuO microspheres. (1) Absorption of aqueous borate buffer solutions (pH 9.0, 80 mmol/L) containing [Ru(bpy)3]Cl2 (1.0 mmol/L), Na2S2O8 (5.0 mmol/L) and CuO microspheres (0.5g/L). (2) Absorption by the above solution after 9 min of irradiation. The concentration of [Ru(bpy)3]2+/sup> decreased by 12.4%.

Fig. S9. (a) Solution containing 1.0 mmol/L of [Ru(bpy)3]2+ and 5.0 mmol/L of Na2S2O8 in pH 8.0, 80 mmol/L NaBi before illumination; (b) Supernatant solution containing 1.0 mmol/L of [Ru(bpy)3]2+ + 5.0 mmol/L of Na2S2O8+0.5 g/L CuO microspheres in pH 8.0, 80 mmol/L NaBi after illumination for 9 min.

Fig. S10. UV-vis spectral changes during the photocatalytic water oxidation with a pH 8.0 buffer. The bottom green line shows the absorption of an aqueous borate buffer solution (pH 8.0, 80 mmol/L) containing [Ru(bpy)3]Cl2 (1.0 mmol/L), Na2S2O8 (5.0 mmol/L) and CuO microspheres (0.5 g/L). Other lines show the UV-vis spectral changes of the green reaction solution obtained by irradiating the initial reaction solution about 30 s.

Fig. S11. Kinetics of O2 formation in the photocatalytic system using different concentrations of [Ru(bpy)3]Cl2. Conditions: LED lamp (λ ≥ 420 nm), 0.5 g/L CuO microspheres, 5.0 mmol/L Na2S2O8, 80 mmol/L sodium borate buffer (initial pH 8.5), total reaction volume 15 mL, overall volume 28 mL, with vigorous agitation using a magnetic stirrer.

Fig. S12. Kinetics of O2 formation in the photocatalytic system use different kinds of buffers. Conditions: LED lamp (λ ≥ 420 nm); catalyst concentration 0.5 g/L, 1.0 mmol/L [Ru(bpy)3]Cl2, 5.0 mmol/L Na2S2O8, 80 mmol/Lsodium borate buffer (initial pH 8.5); total reaction volume 15 mL, overall volume 28 mL, with vigorous agitation using a magnetic stirrer.

Fig. S13. Kinetics of O2 formation in the photocatalytic system use different kinds of buffers. Conditions: LED lamp (λ ≥ 420 nm); catalyst concentration 0.5 g/L, 1.0 mmol/L [Ru(bpy)3]Cl2, 5.0 mmol/L Na2S2O8, total reaction volume 15 mL, overall volume 28 mL, with vigorous agitation using a magnetic stirrer.

Fig. S14. UV-vis spectral changes during the photocatalytic O2 evolution with CuO microspheres. (1) Absorption of aqueous phosphate buffer solutions (pH 8.5, 80 mmol/L) containing [Ru(bpy)3]Cl2 (1.0 mmol/L), Na2S2O8 (5.0 mmol/L) and CuO microspheres (0.5g/L). (2) Absorption by the above solution after 9 min of irradiation. Concentration of [Ru(bpy)3]2+ decreased by 10.2%.

Fig. S15. Kinetics of O2 formation in the photocatalytic system in 80 mmol/L sodium borate buffer (initial pH 8.5), (1)1.0 mmol/L [Ru(bpy)3]Cl2+ 0.5 g/LCuO microspheres; (2)5.0 mmol/L Na2S2O8+0.5 g/L CuO microspheres;(3)1.0 mmol/L [Ru(bpy)3]Cl2+5.0 mmol/L Na2S2O8; (4)1.0 mmol/L [Ru(bpy)3]Cl2+5.0 mmol/L Na2S2O8+0.5 g/LCuO microspheres. Conditions: LED lamp (λ ≥ 420 nm), total reaction volume 15 mL, overall volume 28 mL, with vigorous agitation using a magnetic stirrer.

Fig. S16. Light control experiments of the photochemical water oxidation in 80 mmol/L sodium borate buffer (initial pH 8.5), 1.0 mmol/L [Ru(bpy)3]Cl2 + 5.0 mmol/L Na2S2O8 + 0.5 g/L CuO microspheres.

Fig. S17. XRD patterns of fresh (1) and recovered (2) CuO microspheres.

Fig. S18. FT-IR of fresh (1) and recovered (2) CuO microspheres.

Fig. S19. TEM images of fresh (a) and recovered (b) CuO microspheres.

Fig. S20. Raman spectra of fresh (1) and recovered (2) CuO microspheres excited by 514.5 nm laser.

Fig. S21. Solution containing 1.0 mmol/L of [Ru(bpy)3]2+ + 5.0 mmol/L of Na2S2O8 + 0.5 g/LCu(NO3)2·3H2O in pH 8.5, 80 mmol/L NaBi after illumination for 9 min.