Efficient photocatalytic carbon monoxide production from ammonia and carbon dioxide by the aid of artificial photosynthesis

Abstract

Ammonium bicarbonate (NH₄HCO₃) was generated by the absorption of carbon dioxide (CO₂) into an aqueous solution of ammonia (NH₃). NH₄HCO₃ was successfully used to achieve highly efficient photocatalytic conversion of CO₂ to carbon monoxide (CO). NH₃ and/or ammonium ions (NH₄⁺) derived from NH₄HCO₃ in aqueous solution were decomposed into nitrogen (N₂) and hydrogen (H₂). Stoichiometric amounts of the N₂ oxidation product and the CO and H₂ reduction products were generated when the photocatalytic reaction was carried out in aqueous NH₄HCO₃ solution. NH₃ and/or NH₄⁺ functioned as electron donors in the photocatalytic conversion of CO₂ to CO. A CO formation rate of 0.5 mmol h⁻¹ was obtained using 500 mg of catalyst (approximately 7500 ppm) in ambient conditions (303 K, 101.3 kPa). Our results demonstrated that NH₄HCO₃ is a novel inorganic sacrificial reagent, which can be used to increase the efficiency of photocatalytic CO production to achieve one step CO₂ capture, storage and conversion.

---

Introduction

The production of chemical feedstocks and hydrocarbon fuels from CO₂ is a promising approach to alleviate the global energy crisis and global warming.¹ Conversion of CO₂ to CO using clean and renewable solar energy is the first step to store energy in chemicals because CO can be further converted into other highly valuable chemicals using the Fischer–Tropsch process.² A variety of heterogeneous and homogeneous photocatalysts have been reported to achieve the conversion of CO₂ to CO.^3–5 However, the formation rate of CO has been limited to a few tens of μmol h⁻¹ or hundreds of μmol h⁻¹ g⁻¹ because of the high energy barrier to CO₂ reduction and inefficient light utilization.^6,7 Furthermore, CO₂ is not easily adsorbed onto catalytic surfaces nor activated by photoirradiation because of its high thermodynamic stability. This further reduces the efficiency of the photocatalytic conversion of CO₂.

Water (H₂O) is widely used as an electron donor in the photocatalytic conversion of CO₂ to CO.^7–12 However, the overall water splitting into H₂ and O₂ is more thermodynamically favorable than the reduction of CO₂ in aqueous solution. Hence, CO₂ reduction competes with overall water splitting. Moreover, the solubility of CO₂ in pure H₂O is only 0.033 mol L⁻¹ (at 298 K and 101.3 kPa),¹³ which further limits the efficiency of CO₂ conversion by H₂O using heterogeneous photocatalysts. Therefore, it would be meaningful to find a readily available, highly efficient, and abundant in nature and industries electron donor (sacrificial reagent) other than water for the photocatalytic conversion of CO₂. NH₃ and NH₄⁺ in aqueous solution can be readily oxidized to N₂, NO₂⁻, and NO₃⁻ using a photocatalyst.^14–17 The decomposition of aqueous NH₃ to H₂ and N₂ requires a standard Gibbs free energy change ΔG° of 18 kJ mol⁻¹ (eqn (1)).¹⁸ This is significantly smaller than that required for the decomposition of H₂O to H₂ and O₂ (237 kJ mol⁻¹; eqn (2)).

NH₃(aq) → [/]H₂(g) + ½N₂(g) ΔG° = 18 kJ mol⁻¹ (1)

H₂O(l) → H₂(g) + ½O₂(g) ΔG° = 237 kJ mol⁻¹ (2)

Because the photocatalytic oxidation of NH₃ and NH₄⁺ is significantly more favorable than the oxidation of H₂O to O₂,¹⁸ it is possible to use NH₃ and NH₄⁺ as electron donors in the photocatalytic conversion of CO₂. Moreover, NH₃ has been considered for use as an efficient post-combustion CO₂ capture and storage (CCS) reagent because of its high absorption efficiency and loading capacity.¹⁹ The absorption and capture of CO₂ by an aqueous solution of NH₃ results in the formation of NH₄HCO₃.²⁰ Other basic species, such as NaHCO₃ and KHCO₃, have been used to increase the solubility of CO₂ in aqueous solutions.^21,22 Previous reports have suggested that dissolved CO₂, rather than bicarbonate (HCO₃⁻) or carbonate (CO₃²⁻) ions, is the active species in the reduction of CO₂.^23,24 Correspondingly, the conversion of CO₂ and/or the selectivity toward CO evolution have been significantly enhanced by the presence of bases in both photocatalytic (PC) and photoelectrochemical (PEC) cell systems.^25,26 In the present study, we designed the use of NH₄HCO₃ for the efficient photocatalytic conversion of CO₂ to CO in H₂O.

Results and discussion

Flux-mediated crystal growth method shows the advantage of the synthetic control over particle sizes, morphologies, and surface features comparing with that of solid-state reaction method (SSR).²⁷ Modification of these features as a function of flux conditions have been reported to show significant enhancements in both water splitting and CO₂ photoreduction.^11,27,28 Sr₂KTa₅O₁₅ has been reported to show good activity and selectivity toward CO evolution when used as a photocatalyst in the conversion of CO₂ by H₂O in our previous work.²⁹ In the present study, Sr₂KTa₅O₁₅ was fabricated by a modified flux method, using a mixture of NaCl and KCl as the flux. The resultant catalyst was confirmed to have tetragonal tungsten bronze (TTB) structure (Fig. S1A†), and its real chemical formula was determined to be Sr_1.6K_0.35Na_1.45Ta₅O₁₅ using ICP-OES. Its morphology was observed by SEM, and was found to consist of a mixture of nanorods and nanoparticles (Fig. S1B†).

Fig. 1 shows the time courses of the photocatalytic conversion of CO₂ in H₂O and aqueous solutions of NaHCO₃ and NH₄HCO₃. In pure H₂O, only 16.8 μmol of CO was evolved after 5 h of photoirradiation (Fig. 1A), and the main reduction product was H₂ (139.0 μmol). These results were consistent with previous reports.^25,29 In this system, overall water splitting proceeded more readily than CO₂ reduction, resulting in the generation of H₂ as the major product, rather than CO. The amount of CO evolved in 0.1 M aqueous NaHCO₃ solution after 5 h of photoirradiation (448.7 μmol) was 26.7 times higher than that evolved in pure H₂O (Fig. 1B). However, the formation of H₂ (94.7 μmol) was not significantly affected by NaHCO₃. Thus, NaHCO₃ greatly enhanced the conversion of CO₂ to CO without affecting the water splitting process.²⁵ In both pure H₂O and 0.1 M aqueous NaHCO₃, stoichiometric amounts of O₂ were evolved continuously during the reaction, implying that H₂O functioned as an electron donor in the reduction of CO₂. Moreover, the evolution of CO increased dramatically in 0.1 M aqueous NH₄HCO₃ solution; 1600 μmol (1.6 mmol) of CO was evolved after 5 h of photoirradiation (Fig. 1C). This is 94.2 times greater than the amount evolved in pure H₂O. The selectivity of the reaction toward CO evolution was calculated and the details were shown in ESI.† The selectivity toward CO evolution in 0.1 M aqueous NH₄HCO₃ (86.2%) was similar to that in aqueous NaHCO₃ (82.5%). The production of gaseous products was negligible in blank tests conducted without either a catalyst or photoirradiation (Fig. S2A and B†). Thus, both are necessary for the photocatalytic conversion of CO₂ to CO to proceed. Without Ag cocatalyst, H₂ was formed as main product (Fig. S2C†), both of N₂ and O₂ were detected as oxidation products, however, the amount of these gases was far beyond the stoichiometric amount. Tiny amount of CO was formed after 5 hour photoirradiation (14.9 μmol). Ag cocatalysts were important in photocatalytic conversation of CO₂ to CO, which is thought to be the active sites. H₂, CO, and N₂ were obtained without a continuous CO₂ flow (Fig. S2D†). However, H₂ was generated as a major product, suggesting a very low selectivity toward CO evolution (less than 30%). This suggested that CO₂ presence significantly increases the selectivity of the photocatalytic conversion of CO₂ toward CO evolution in NH₄HCO₃ solution. NH₄HCO₃ can be formed directly by the absorption of CO₂ in an aqueous solution of NH₃; wherein H₂ and CO can be produced from CO₂ and NH₃via artificial photosynthesis. Thus, our designed system can achieve carbon capture and utilization (CCU) in a single process.

Fig. 1 Time courses of CO (circle), O₂ (square), N₂ (lozenge), and H₂ (triangle) evolutions during the photocatalytic conversion of CO₂ over Ag-modified Sr_1.6K_0.35Na_1.45Ta₅O₁₅. Amount of catalyst: 0.5 g; cocatalyst loading: 1.0 wt% Ag; light source: 400 W high-pressure Hg lamp; water volume: 1.0 L; CO₂ flow rate: 30 mL min⁻¹; additive: (A) none, (B) 0.1 M NaHCO₃, or (C) 0.1 M NH₄HCO₃.

Fig. 2 Gas chromatogram and mass spectra (m/z = 28 and 29) obtained during the photocatalytic conversion of ¹³CO₂ using Ag-modified Sr_1.6K_0.35Na_1.45Ta₅O₁₅. Amount of catalyst: 0.5 g; cocatalyst loading: 5.0 wt% Ag; light source: 400 W high-pressure Hg lamp; water volume: 1.0 L; CO₂ flow rate: 30 mL min⁻¹; additive: 0.5 M NH₄HCO₃.

N₂, rather than O₂, was generated as the oxidation product during photoirradiation in the presence of NH₄HCO₃ (Fig. 1C). This demonstrated that H₂O did not function as an electron donor in this system. Instead, NH₃ and/or NH₄⁺ functioned as electron donors, because the ΔG° of NH₃(aq) oxidation (18 kJ mol⁻¹) is significantly lower than that of water oxidation (237 kJ mol⁻¹). Analysis of the liquid phase showed that neither NO₂⁻ nor NO₃⁻ were present during photoirradiation (Fig. S3†). Other gaseous NO_x products, such as N₂O and NO, were not detected by gas chromatography (GC). These results indicated that NH₃ and/or NH₄⁺ were oxidized only to N₂ in this photocatalytic system. Hence, by using NH₄HCO₃, we succeeded in controlling the oxidation product, in addition to enhancing the conversion of CO₂.

The ratio of electrons to holes consumed in the photocatalytic conversion of CO₂ was calculated to be 2.0 after 1 h of photoirradiation (Fig. S4†). Given that the total number of electrons generated must be the same as the number of holes, this ratio indicates that significantly more electrons were consumed than holes in the initial stages of photoirradiation. We noted that the state of Ag was changed from metallic to Ag⁺ on the surface of catalyst measured by XPS (Fig. S5†), however, it might be not the main reason for the excess of electron consumption. We calculated that if all Ag⁰ was changed to Ag⁺ in the first hour, the consumed holes were still only 469 μmol, which was much less than the consumed electrons (770 μmol). NH₄⁺ can be reduced to NH₃ and ˙H by photogenerated electrons. Hydrazine (N₂H₄) has been determined to be an intermediate species in the photocatalytic decomposition of NH₃ and/or NH₄⁺ to H₂ and N₂ using Pt/TiO₂.¹⁴ Stoichiometric amounts of products, including H₂ and N₂, were not obtained in the initial stages of photoirradiation due to the formation of hydrazine. In our system, it is also possible to form hydrazine at the beginning, however, hydrazine is reported to be reacted with CO₂ to form zwitterionic intermediate and carbamate-type species,³⁰ which made the detection of intermediate oxidation species much more difficult. Nevertheless, stoichiometric amounts of products were obtained after 2 h of photoirradiation (Fig. S4†), indicating that the total decomposition of NH₃ and/or NH₄⁺ occurred sooner.

The above results demonstrate that the highly efficient photocatalytic conversion of CO₂ to CO was achieved in our system. The stoichiometric amounts of H₂, N₂ and CO generated indicated that NH₃ and/or NH₄⁺ functioned as electron donors in the photocatalytic conversion of CO₂. Significantly greater photocatalytic activity was obtained using NH₃ and/or NH₄⁺, compared to reactions using H₂O as an electron donor under the same conditions. NH₃ and/or NH₄⁺ are suitable for use in practical applications because NH₃ is industrially produced in large quantities. Furthermore, in our photocatalytic system, NH₃ and/or NH₄⁺ can be completely decomposed to N₂, which is an inert and non-toxic gas.

Table 1 shows the effects of NH₄HCO₃ concentration on the photocatalytic conversion of CO₂. In pure H₂O, overall water splitting proceeded as the dominant reaction. Hence, the evolution of CO was negligible (entry 1). When the photocatalytic reaction was carried out in 0.01 M aqueous NH₄HCO₃ the production of H₂ resulting from water splitting was dramatically suppressed (entry 2). Because the oxidation of NH₃ and/or NH₄⁺ to N₂ proceeds more readily than the oxidation of H₂O to O₂, the formation rate of O₂ in 0.01 M aqueous NH₄HCO₃ was less than half that in pure H₂O. Even low concentrations of NH₄HCO₃ (0.01 M) significantly increased the formation rate of CO, indicating that the presence of NH₄HCO₃ is vital to achieving high photocatalytic activity. NH₄HCO₃ can also be used to increase the pH of the reaction solution, to offset the decrease in pH caused by the dissolution of CO₂. With CO₂ flowing, the pH of the reaction solution based on pure H₂O was 3.95, which increased to 5.88 with the addition of 0.01 M NH₄HCO₃. Increasing the pH also increases the amount of CO₂ that can be dissolved in the reaction solution.³¹ Generally, the formation rate of CO increases with increasing pH, because the reaction rate largely depends on the concentration of substrate. Therefore, the addition of NH₄HCO₃ contributed to the efficient conversion of CO₂ and the good selectivity toward CO evolution. Increasing the concentration of NH₄HCO₃ from 0.01 M to 0.05 M completely suppressed the overall water splitting reaction, since only N₂ was generated as an oxidation product (entry 3). The formation rate of CO increased with the concentration of NH₄HCO₃. Increasing the NH₄HCO₃ concentration from 0.1 M to 1.0 M increased the formation rate of CO to 550.7 μmol h⁻¹ except the selectivity toward CO evolution decreased slightly, from 86.1% to 65.5% (entry 4 to 7). As previously discussed, NH₄HCO₃ can be synthesized by flowing CO₂ through an aqueous solution of NH₃. To determine whether NH₃ functions as an electron donor under a flow of CO₂, we carried out the photocatalytic conversion of CO₂ in an aqueous solution of NH₃ (entry 8). The formation rate of CO was 547.2 μmol h⁻¹ in ca. 0.5 M aqueous NH₃, indicating that NH₃ functions efficiently as an electron donor under these conditions. The ratio of photogenerated electrons to holes (e⁻/h⁺) was estimated to be around 1.0 in reactions with high concentrations of aqueous NH₄HCO₃ after 5 h of photoirradiation. This further supports the hypothesis that NH₃ and/or NH₄⁺ function as effective electron donors during the photocatalytic conversion of CO₂.

Table 1 Photocatalytic conversion of CO₂ over Ag-modified Sr_1.6K_0.35Na_1.45Ta₅O₁₅ with different additive concentrations. Amount of catalyst: 0.5 g; cocatalyst loading: 1.0 wt% Ag; light source: 400 W high-pressure Hg lamp; water volume: 1.0 L; CO₂ flow rate: 30 mL min⁻¹

Entry	NH4HCO3^a/M	Formation rate^b/μmol h^−1				Selec.^c (%)	e^−/h^+^d
		H2	O2	N2	CO
a Additive concentration used for CO2 conversion. b Formation rate after 5 h of irradiation. c Selectivity toward CO evolution. d Ratio of consumed electrons to holes after 5 h of irradiation. e 0.5 M aqueous NH3 solution was used as the additive, instead of NH4HCO3.
1	0	35.9	16.3	Trace	3.6	9.2	1.21
2	0.01	16.9	7.0	12.3	54.5	76.3	1.17
3	0.05	23.8	Trace	42.3	146.7	86.1	1.34
4	0.1	48.4	Trace	94.3	270.4	84.8	1.13
5	0.5	119.8	Trace	193.6	512.9	81.1	1.09
6	0.8	175.4	Trace	213.3	520.0	74.8	1.09
7	1.0	290.1	Trace	258.1	550.7	65.5	1.09
8^e	—	235.0	Trace	244.9	547.2	70.0	1.06

---

To confirm that CO evolution originated from CO₂ introduced in the gas phase, rather than from carbon contaminants, we conducted an isotopic labeling experiment. Fig. 2 shows mass spectra (m/z = 28 and 29) obtained during the photocatalytic conversion of ¹³CO₂ in 0.5 M aqueous NH₄HCO₃ over Ag-modified Sr_1.6K_0.35Na_1.45Ta₅O₁₅ after 0.5 h of photoirradiation. Gaseous samples were introduced into a mass spectrometer (MS) after separation by thermal conductivity detector-gas chromatography (TCD-GC). CO was observed in both the gas chromatogram and the mass spectra. The peak positions in the mass spectra were consistent with those in the chromatogram. The major product was ¹³CO, rather than ¹²CO. The presence of a small amount of ¹²CO may be due to the direct decomposition of NH₄HCO₃ since the decomposition of NH₄HCO₃ was observed in samples without a CO₂ flow (Fig. S2D†). The amount of ¹³CO estimated by mass spectrometry was approximately equal to the amount of CO determined using a flame ionization detector (FID-GC) (Fig. S6†). These results demonstrate that CO was predominantly generated from CO₂ introduced in the gas phase, rather than from other carbon resources.

The recycle test was also performed to confirm the stability and durability of our catalyst and system using the Sr_1.6K_0.35Na_1.45Ta₅O₁₅ photocatalyst repeatedly for three times under the same conditions (Fig. S7†). In the second cycle, there is a slight loss by ca. 10% of CO evolution activity during 5 h photoirradiation as compared to the first run, however, the evolution of H₂ showed no obvious changes. The slight loss of activity should be due to the change of Ag cocatalyst (Fig. S5†).²⁹ The photocatalytic activity of CO, N₂, and H₂ were stabilized at ca. 0.5, 0.19 and 0.07 mmol h⁻¹, respectively, during the second and third runs. The structure of catalyst itself was stable during the three cycles (Fig. S8†). These results suggested that the photocatalyst and the system exhibit favorable stability to form CO, N₂, and H₂ during the photocatalytic conversion of CO₂.

To confirm the versatility of NH₄HCO₃ as a general electron donor in photocatalytic reactions, we carried out the photocatalytic conversion of CO₂ in aqueous NH₄HCO₃ solution over 4 types of photocatalysts. All these photocatalysts have been already reported to show good activity and high selectivity toward CO evolution in the photocatalytic conversion of CO₂ using H₂O as an electron donor.^25,32–34 As shown in Table 2, all the photocatalysts showed good activity for conversion of CO₂ and high selectivity toward CO evolution. The activities of the photocatalysts were significantly increased in aqueous NH₄HCO₃ solution, compared with their reported activities in pure H₂O or aqueous NaHCO₃.^32–34 N₂ was detected as the only oxidation product and the e⁻/h⁺ ratio was approximately equal to 1.0. These results indicated that NH₃ and/or NH₄⁺ was easily decomposed to N₂ gas by the photocatalysts tested.

Table 2 Photocatalytic conversion of CO₂ over Ag-modified catalysts in aqueous NH₄HCO₃ solution. Amount of catalyst: 0.5 g; cocatalyst loading: 5.0 wt% Ag; light source: 400 W high-pressure Hg lamp; water volume: 0.95 L; CO₂ flow rate: 30 mL min⁻¹; additive: 0.5 M NH₄HCO₃

Entry	Catalyst	Formation rate^a/μmol h^−1			Selec.^b (%)	e^−/h^+^c
		H2	N2	CO
a Formation rate after 5 h of irradiation. O2 was not detected in any of the samples. b Selectivity toward CO evolution. c Ratio of consumed electrons to holes after 5 h of irradiation.
1	ZnGa2O4/Ga2O3	125.2	191.4	532.0	80.9	1.14
2	ZnGa2O4	39.4	94.2	305.4	88.6	1.22
3	La2Ti2O7	5.9	17.4	41.6	87.6	0.91
4	SrO/Ta2O5	2.71	11.5	42.9	94.1	1.32

---

Conclusions

We designed a highly efficient process for the photocatalytic conversion of CO₂ to CO in aqueous NH₄HCO₃ solution. The stoichiometric formation of CO, H₂, and N₂ indicated that NH₃ and/or NH₄⁺ were consumed as electron donors, instead of H₂O. NH₄HCO₃ was determined to be an effective electron donor for the photocatalytic conversion of CO₂, whereby CO₂ can be captured, stored, and efficiently converted into CO. This novel inorganic additive is suitable for use in carbon capture and utilization process. This new process is a promising way to control the conversion of CO₂ to CO and efficiently produce H₂ and CO.

Acknowledgements

This study was partially supported by a Grant-in-Aid for Scientific Research on Innovative Areas “All Nippon Artificial Photosynthesis Project for Living Earth” (No. 2406) from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan, the Precursory Research for Embryonic Science and Technology (PRESTO), supported by the Japan Science and Technology Agency (JST), and the Program for Elements Strategy Initiative for Catalysts & Batteries (ESICB), commissioned by the MEXT of Japan. Zeai Huang thanks the State Scholarship of China Scholarship Council, affiliated with the Ministry of Education of the P. R. China.

Notes and references

1. S. N. Habisreutinger, L. Schmidt-Mende and J. K. Stolarczyk, Angew. Chem., Int. Ed., 2013, 52, 7372–7408.
2. G. Centi, E. A. Quadrelli and S. Perathoner, Energy Environ. Sci., 2013, 6, 1711–1731.
3. K. Teramura, S. Iguchi, Y. Mizuno, T. Shishido and T. Tanaka, Angew. Chem., 2012, 124, 8132–8135.
4. R. Kuriki, K. Sekizawa, O. Ishitani and K. Maeda, Angew. Chem., Int. Ed., 2015, 54, 2406–2409.
5. R. Kuriki, H. Matsunaga, T. Nakashima, K. Wada, A. Yamakata, O. Ishitani and K. Maeda, J. Am. Chem. Soc., 2016, 138, 5159–5170.
6. S. Navalón, A. Dhakshinamoorthy, M. Álvaro and H. Garcia, ChemSusChem, 2013, 6, 562–577.
7. X. Chang, T. Wang and J. Gong, Energy Environ. Sci., 2016, 9, 2177–2196.
8. G. Liu, S. Xie, Q. Zhang, Z. Tian and Y. Wang, Chem. Commun., 2015, 51, 13654–13657.
9. K. Iizuka, T. Wato, Y. Miseki, K. Saito and A. Kudo, J. Am. Chem. Soc., 2011, 133, 20863–20868.
10. M. Yamamoto, T. Yoshida, N. Yamamoto, T. Nomoto, Y. Yamamoto, S. Yagi and H. Yoshida, J. Mater. Chem. A, 2015, 3, 16810–16816.
11. H. Yoshida, L. Zhang, M. Sato, T. Morikawa, T. Kajino, T. Sekito, S. Matsumoto and H. Hirata, Catal. Today, 2015, 251, 132–139.
12. H. Nakanishi, K. Iizuka, T. Takayama, A. Iwase and A. Kudo, ChemSusChem, 2017, 10, 112–118.
13. K. Hara, A. Kudo, T. Sakata and M. Watanabe, J. Electrochem. Soc., 1995, 142, L57–L59.
14. H. Yuzawa, T. Mori, H. Itoh and H. Yoshida, J. Phys. Chem. C, 2012, 116, 4126–4136.
15. H. Wang, Y. Su, H. Zhao, H. Yu, S. Chen, Y. Zhang and X. Quan, Environ. Sci. Technol., 2014, 48, 11984–11990.
16. X. Zhu, S. R. Castleberry, M. A. Nanny and E. C. Butler, Environ. Sci. Technol., 2005, 39, 3784–3791.
17. S. Yamazoe, Y. Hitomi, T. Shishido and T. Tanaka, Appl. Catal., B, 2008, 82, 67–76.
18. H. Kominami, H. Nishimune, Y. Ohta, Y. Arakawa and T. Inaba, Appl. Catal., B, 2012, 111, 297–302.
19. B. T. Zhao, Y. X. Su, W. W. Tao, L. L. Li and Y. C. Peng, Int. J. Greenhouse Gas Control, 2012, 9, 355–371.
20. X. Wang, W. Conway, D. Fernandes, G. Lawrance, R. Burns, G. Puxty and M. Maeder, J. Phys. Chem. A, 2011, 115, 6405–6412.
21. S. F. Shen, Y. N. Yang and S. F. Ren, Fluid Phase Equilib., 2014, 367, 38–44.
22. H. Zhong, K. Fujii, Y. Nakano and F. Jin, J. Phys. Chem. C, 2015, 119, 55–61.
23. B. Kumar, M. Llorente, J. Froehlich, T. Dang, A. Sathrum and C. P. Kubiak, Annu. Rev. Phys. Chem., 2012, 63, 541–569.
24. Y. Hori, A. Murata and R. Takahashi, J. Chem. Soc., Faraday Trans. 1, 1989, 85, 2309–2326.
25. K. Teramura, Z. Wang, S. Hosokawa, Y. Sakata and T. Tanaka, Chem.–Eur. J., 2014, 20, 9906–9909.
26. E. E. Barton, D. M. Rampulla and A. B. Bocarsly, J. Am. Chem. Soc., 2008, 130, 6342–6344.
27. J. Boltersdorf, N. King and P. A. Maggard, CrystEngComm, 2015, 17, 2225–2241.
28. Y. Ham, T. Hisatomi, Y. Goto, Y. Moriya, Y. Sakata, A. Yamakata, J. Kubota and K. Domen, J. Mater. Chem. A, 2016, 4, 3027–3033.
29. Z. Huang, K. Teramura, S. Hosokawa and T. Tanaka, Appl. Catal., B, 2016, 199, 272–281.
30. K. H. Lee, B. Lee, J. H. Lee, J. K. You, K. T. Park, I. H. Baek and N. H. Hur, Int. J. Greenhouse Gas Control, 2014, 29, 256–262.
31. K. Teramura, K. Hori, Y. Terao, Z. Huang, S. Iguchi, Z. Wang, H. Asakura, S. Hosokawa and T. Tanaka, J. Phys. Chem. C, 2017, 121, 8711–8721.
32. Z. Wang, K. Teramura, S. Hosokawa and T. Tanaka, J. Mater. Chem. A, 2015, 3, 11313–11319.
33. Z. Wang, K. Teramura, S. Hosokawa and T. Tanaka, Appl. Catal., B, 2015, 163, 241–247.
34. K. Teramura, H. Tatsumi, Z. Wang, S. Hosokawa and T. Tanaka, Bull. Chem. Soc. Jpn., 2015, 88, 431–437.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, calculations and characterizations. See DOI: 10.1039/c7sc01851g

---