Organophotocatalysis system of p/n bilayers for wide visible-light-induced molecular hydrogen evolution

Abstract

We developed a photocatalysis system composed of two types of organic p/n bilayers for the evolution of H₂. When a bilayer of 29H,31H-phthalocyanine (H₂Pc, a p-type semiconductor)/fullerene (C₆₀, an n-type semiconductor) was applied to a photocatalyst by loading a Pt co-catalyst onto the C₆₀ surface, the Pt-coated C₆₀ surface induced H₂ evolution, concurrently involving photocatalytic oxidation at H₂Pc in the other bilayer. The photocatalytic reaction can occur along with photophysical events in the p/n interior. In order to draw out an efficient photocatalysis, the factors affecting H₂ evolution were investigated in terms of the intensity of light irradiated, thickness of the bilayers, and the amount of Pt co-catalyst deposited. The organophotocatalysis system demonstrates a novel and advanced method of photocatalytic H₂ evolution utilizing energy over the entire range of visible light.

---

Introduction

Molecular hydrogen (H₂) is widely recognized as a potential clean energy source, and recently fuel cells have become practical devices for energy generation. Some fuel cells use H₂, but a clean procedure for H₂ production has not been established. Unless this problem is solved, a widespread public energy system that relies on the production and consumption of H₂ cannot be created. Water photolysis is the most ideal method for producing H₂. Since a photoelectrochemical approach for water splitting featuring a UV-responsive TiO₂ photoanode was reported by Honda and Fujishima,¹ the fabrication of inorganic semiconductor-based photocatalysts responsive to visible light has been passively conducted by means of band engineering (e.g. doping, solid solution, and the formation of a valence band alternative to O 2p orbitals for a more negative potential position).^2–5 Although there is high demand for a visible-light-responsive photocatalyst that can achieve an efficient reaction of water, few examples of photocatalysts responsive to visible light with λ > 600 nm have been developed;^6–9 that is, there are almost no instances of a wide visible-light-induced reaction of water. It should also be considered that the visible light response of the band-engineered photocatalysts may trade off its activities. Therefore, a new approach is needed to achieve efficient photocatalysis that utilizes a broad region of the visible light spectrum for facilitating the reaction of water in order to advance the development of photochemical energy conversion.

We have previously demonstrated the novel function of organic p/n bilayers as photoelectrodes, particularly in the water phase,^10–14 in which the bilayer can induce a chemical reaction at the solid/water interface along with photophysical events (i.e. light absorption, charge separation at the p/n interface, and carrier conduction) in its interior. Moreover, in the photoelectrodes of the p/n bilayer, the photoanodic and photocathodic effects can occur at the surface of the p-type and n-type conductors, respectively. Therefore, these studies prompted us to use the organic p/n bilayer as a photocatalyst. Here, we describe a photocatalysis system composed of two types of organic p/n bilayers set in a cell with twin compartments (Scheme 1), one of which is composed of 29H,31H-phthalocyanine (H₂Pc, a p-type semiconductor)/fullerene (C₆₀, an n-type semiconductor) for reduction and the other is a bilayer of 3,4,9,10-perylenetetracarboxylic-bis-benzimidazole (PTCBI, an n-type semiconductor)/H₂Pc for oxidation. Although organic dyes have often been applied for the sensitisation of inorganic semiconductors,^9,15–17 this system presents a novel example of light-induced H₂ evolution by organophotocatalysis using dyes, and in particular, it utilizes the entire visible light region.

Scheme 1 Schematic illustration of the photocatalysis system of H₂Pc/C₆₀ and PTCBI/H₂Pc.

Experimental

Chemicals

Pure C₆₀ (>99.5%) was purchased from Tokyo Kasei Kogyo Co., Ltd. and used as received. PTCBI was synthesized and purified according to a previously described procedure.¹⁸ Commercially available H₂Pc (Tokyo Kasei Co., Ltd.) was purified by sublimation prior to use. During sublimation, the exterior of the vessel was thermally controlled (temperature, 510 °C; pressure within the vessel, ∼1.0 × 10⁻² Pa). 2-Mercaptoethanol (RS⁻) was obtained from Kanto Chemical Co., Inc. The indium–tin oxide (ITO)-coated glass plate (sheet resistance, 8 Ω cm⁻²; transmittance, >85%; ITO thickness, 174 nm) was obtained from Asahi Glass Co., Ltd.

Film preparation

The organic bilayers were prepared by vapour deposition (pressure, <1.0 × 10⁻³ Pa; deposition speed, ca. 0.03 nm s⁻¹), for which an ITO-coated glass plate was used as base material. The bilayer of H₂Pc/C₆₀ comprised H₂Pc coated on ITO, and C₆₀ coated on top of the H₂Pc layer (denoted as ITO/H₂Pc/C₆₀). To generate the PTCBI/H₂Pc bilayer, PTCBI was first coated on ITO, followed by a coating of H₂Pc (denoted as ITO/PTCBI/H₂Pc). During vapour deposition, the temperature at the ITO plate was not controlled. Absorption spectral measurements were conducted using a Hitachi U-2010 spectrophotometer. The resulting absorption spectra of C₆₀,¹⁹ PTCBI,²⁰ and H₂Pc (α-phase)^20,21 were identical to those reported previously, and their absorption coefficients were used to determine the thickness of the film.^13,14

The deposition of Pt on ITO/H₂Pc/C₆₀ was carried out as follows: an electrochemical glass cell with a single compartment was equipped with a modified ITO working electrode (effective area, 1 cm²), a spiral Pt counter electrode, and an Ag/AgCl (in saturated KCl electrolyte solution) reference electrode. ITO/H₂Pc/C₆₀ was photocathodically polarized from +0.4 V to −0.2 V in acidic solution (pH = 2) containing 5.0 × 10⁻⁴ mol dm⁻³ H₂PtCl₆·6H₂O. The product was denoted ITO/H₂Pc/C₆₀–Pt, wherein the amount of deposited Pt was controlled through the amount of charge passed (typically, 2.0 × 10⁻² C). This was performed using a potentiostat (Hokuto Denko, HA-301) with a function generator (Hokuto Denko, HB-104), a coulomb meter (Hokuto Denko, HF-201), and an X–Y recorder (GRAPHTEC, WX-4000) under illumination (light source, a halogen lamp; light intensity, ca. 100 mW cm⁻²). Irradiation was carried out from the back side of ITO-coated face.

The surface roughness of both C₆₀ in the H₂Pc/C₆₀ bilayer (i) and Pt-coated C₆₀ in the H₂Pc/C₆₀–Pt photocatalyst (ii) was measured by a laser microscope (Keyence, VK-9510). The roughness factor, calculated from the ratio of real surface area to geometrical area, was estimated to be ca. 2 and 2.3 in systems (i) and (ii), respectively.

Measurements

A cell with twin compartments separated by a salt bridge was used for the photocatalytic experiments (see Scheme 1). ITO/H₂Pc/C₆₀–Pt and ITO/PTCBI/H₂Pc were placed in Compartments A and B, respectively, and the two photocatalyst devices were connected with a lead wire. The effective area of each photocatalyst device was 1 cm² (1 cm × 1 cm). The photocatalytic study was carried out under an Ar atmosphere, where ITO/H₂Pc/C₆₀–Pt was immersed in a phosphoric acid solution (pH = 2) and ITO/PTCBI/H₂Pc in an alkaline (KOH) solution containing a known concentration of thiol (pH = 11). For preparing the salt bridge, agar (1.3 g) and KNO₃ (4.74 g) were first dissolved in hot water (1.0 × 10⁻² dm³), and then the mixture was allowed to flow into the bridging region of the cell (Scheme 1) and to solidify at room temperature. A halogen lamp was usually used as the light source, and irradiation was carried out from the Pt-coated C₆₀ surface of the sample. For irradiation with monochromatic light, the lamp was used with a monochromator (Soma Optics, Ltd., S-10). A piece of shading board was placed between the two compartments to prevent the mixing of the light irradiated from distinct sources. The light intensity was measured using a power meter (type 3A from Ophir Japan, Ltd.). The produced H₂ was analyzed using a thermal conductivity detector gas chromatograph (Shimadzu, GC-8A) equipped with a molecular sieve 5 Å column and Ar carrier gas. Quantification of H₂ was performed using a chromatogram analyzer (Shimadzu, C-R8A) equipped with the chromatograph.

Calculation methods

The external quantum efficiency (EQE) of H₂ evolution was estimated as follows:

EQE = (N_A × M)/(I × A) (1)

where N_A is Avogadro's number, M is the total number of reacted electrons in H₂ evolution, I is the number of incident photons at a given wavelength, and A is the total area of the photocatalyst films employed [in the present system, 1 cm² (oxidation part) + 1 cm² (reduction part)], respectively.

Results and discussion

The amount of H₂ evolved in the photocatalysis system of ITO/H₂Pc/C₆₀ and ITO/PTCBI/H₂Pc was determined and compared with four different control systems (Table 1). Under the reported conditions, the photocatalysis system generated 265.3 μl of H₂ gas (Run 1). As negative controls, when only one of the two photocatalyst devices was irradiated (Runs 2 and 3), Pt was absent from the C₆₀ surface (Run 4), or when no thiol was added in Compartment B (Run 5), virtually no H₂ evolution in Compartment A was observed. Therefore, it appeared that the present photocatalysis system achieved reduction of H⁺ at the Pt-coated C₆₀ surface of ITO/H₂Pc/C₆₀ (i.e. ITO/H₂Pc/C₆₀–Pt) and oxidation of thiol at the H₂Pc/water interface of ITO/PTCBI/H₂Pc.

Table 1 The amount of H₂ evolved in the photocatalysis system of ITO/H₂Pc/C₆₀–Pt (in Compartment A) and ITO/PTCBI/H₂Pc (in Compartment B)

System	H2 amount/μl	Note
a Photocatalysis occurs according to Scheme 1. The light intensity of 25 mW cm^−2 was irradiated for each photocatalyst device, i.e. the sum of the intensities irradiated for ITO/H2Pc/C60–Pt (effective area, 1 cm^2) and ITO/PTCBI/H2Pc (effective area, 1 cm^2) was 50 mW cm^−2. Irradiation was carried out from the Pt-coated C60 surface of the sample in Compartment A and the back side of ITO-coated face in Compartment B. Other conditions were as follows: irradiation time, 3 h; electrolyte solution in Compartment A, a phosphoric acid solution (pH = 2); electrolyte solution in Compartment B, an alkaline solution of 2-mercaptoethanol (concentration = 5 mmol dm^−3, pH = 11); the amount of charge passed during Pt deposition, 2.0 × 10^−2 C; film thickness of the H2Pc/C60 bilayer, H2Pc = 75 nm and C60 = 125 nm and film thickness of the PTCBI/H2Pc bilayer, PTCBI = 300 nm and H2Pc = 60 nm.
Run 1	265.3	Full conditions^a
Run 2	0	Without irradiation for ITO/H2Pc/C60–Pt
Run 3	14.0	Without irradiation for ITO/PTCBI/H2Pc
Run 4	1.0	No deposition of Pt onto ITO/H2Pc/C60
Run 5	0	No addition of the electron donor (i.e. thiol)

---

Based on the dependence of H₂ generation on light intensity (Fig. 1) and thiol concentration (Fig. 2), the conditions for Run 1 are such that the evolution of H₂ is the rate-limiting step. For intensities higher than 2.5 mW cm⁻² in Fig. 1, the amount of H₂ evolved was not proportional to light intensity, indicating that the generation of carriers was not the rate-determining step under these conditions. At concentrations higher than 5 mM (Fig. 2), the amount of H₂ was not proportional to thiol concentration, which indicates that the overall kinetics of the photocatalysis system was not dominated by the oxidation of thiol under these conditions.

Fig. 1 Dependence of the amount of H₂ generated on light intensity. The conditions for Run 1 listed in Table 1 were applied to the photocatalysis system, with the exception that the irradiation time (30 min) and the intensity of white light used to irradiate ITO/H₂Pc/C₆₀–Pt were varied. Light of constant intensity (25 mW cm⁻²) was used to irradiate ITO/PTCBI/H₂Pc.

Fig. 2 Dependence of the amount of H₂ generated on thiol concentration. The conditions for Run 1 in Table 1 were applied to the photocatalysis system, with the exception that the thiol concentration in compartment B was varied.

In addition to light intensity and thiol concentration, the dependence of H₂ generation on the film thickness of H₂Pc/C₆₀ (Fig. 3) and the amount of Pt co-catalyst (Fig. 4) was also considered. By holding the thickness of C₆₀ (i.e. 125 nm) and varying the thickness of H₂Pc, the optimum H₂Pc thickness was found to be 75 nm (Fig. 3a), indicating that a thick layer of H₂Pc was effective for carrier generation. However, in the thicker H₂Pc, an increase in electron–hole concentration can occur. On that occasion, electron–hole recombination becomes more probable with an increase in the concentration, thus resulting in a decrease in H₂ evolution. Similarly, the optimum thickness of the C₆₀ layer was investigated by holding the thickness of H₂Pc constant at 75 nm and varying the C₆₀ thickness (Fig. 3b). This experiment also indicated that a thick layer of C₆₀ was most effective. However, a thicker C₆₀ layer may induce an optical filter effect, disturbing the light absorption of H₂Pc as well as C₆₀ next to the p/n interface, thus leading to lower H₂ evolution. The amount of H₂ generated increased with the amount of Pt deposited up to 0.02 C (Fig. 4). When the loading of the co-catalyst is effective for the rate-limiting reaction, the H₂ amount essentially increases. However, since the number of electrons photogenerated in the H₂Pc/C₆₀–Pt film is constant under the experimental conditions employed in the experiment reported in Fig. 4, an increase in the number of catalysis sites may cause a decrease in the activity per co-catalyst (i.e. decreasing the concentration of electrons at the co-catalyst), thus resulting in a decrease in H₂ generation. Therefore, Fig. 4 may imply that the positive and negative factors for H₂ evolution are balanced, particularly when the charge passed during Pt deposition is 0.04 C. Furthermore, the relatively higher loading of Pt caused a conspicuous depression of the amount of H₂ generated. In consideration of the direction of irradiation (vide supra), an interruption of incident light by Pt may also contribute to the decrease in the amount of H₂ generated. Considering the above-mentioned parameters, the results suggest that Run 1 was conducted under virtually optimal conditions.

Fig. 3 Dependence of the amount of H₂ generated on film thickness of H₂Pc/C₆₀. The conditions for Run 1 in Table 1 were applied to the photocatalysis system, with the exception that the thickness was varied. Concretely, this study was conducted by changing the thickness of only one layer [e.g. H₂Pc (a) or C₆₀ (b)] in the bilayer.

Fig. 4 Dependence of the amount of H₂ generated on that of Pt coated on the C₆₀ surface. The conditions for Run 1 in Table 1 were applied to the photocatalysis system, with the exception that the Pt amount was varied. In all cases studied, the faradaic efficiency for Pt deposition was calculated as ca. 70% (i.e. during photocathodic polarization for depositing Pt, the reduction of H⁺ to H₂ concurrently occurred with ca. 30% of the faradaic efficiency).

Photocatalytic H₂ evolution was also investigated under irradiation with monochromatic light. The EQE of H₂ evolution was estimated for each of the incident wavelengths tested, and the results are shown in Fig. 5. The dependence of EQE on the incident wavelength was rather consistent with the sum of the absorption spectra of the bilayers employed (the spectrum of each bilayer is shown in ESI† (Fig. S1)). Furthermore, note that all visible light at λ < 750 nm contributed to H₂ evolution.

Fig. 5 Dependence of EQE on the incident wavelength for the photocatalysis system and the sum of the absorption spectra of H₂Pc/C₆₀ and PTCBI/H₂Pc. For measuring EQE, the conditions of Run 1 in Table 1 were applied to the photocatalysis system, with the exception that monochromatic light was irradiated for each photocatalyst device. The intensity irradiated for a photocatalyst device was 0.83 mW cm⁻².

A prolonged study was conducted in which the ITO/H₂Pc/C₆₀–Pt photocatalyst device was repeatedly used to test its durable performance for H₂ generation. The conditions for this study were chosen based on the optimal experimental conditions found in the experiments, as shown in Fig. 1–4. Fig. 6 shows the relationship between the amount of H₂ generated and the cycle number, explicitly demonstrating that stable photocatalysis for H₂ evolution at ITO/H₂Pc/C₆₀–Pt occurred even after several cycles.

Fig. 6 A prolonged study of photocatalytic H₂ evolution in the present system. The conditions of Run 1 in Table 1 were applied to the photocatalysis system, with the exception that the charge passed during Pt deposition (4.0 × 10⁻² C) and the irradiation time (6 h in one cycle) were varied. In every cycle, the electrolyte solutions in both compartments were replaced with fresh solutions.

The mechanism of photocatalytic H₂ evolution at ITO/H₂Pc/C₆₀–Pt is suggested as follows: the fundamental characteristics of ITO/PTCBI/H₂Pc and ITO/H₂Pc/C₆₀–Pt have been characterized by means of photoelectrochemistry,^13,14 through which we confirmed that each bilayer can generate photovoltage in the p/n interior even when applied to a photoelectrode in the water phase. This supports the suggestion that the photophysical events in the bilayer are almost similar to those of the corresponding photovoltaic cell in the dry state.²² However, particularly for ITO/PTCBI/H₂Pc in the water phase, carriers can be generated only from the exciton of PTCBI through charge separation at the p/n interface,^10,13 which was also supported by the previous study of photoluminescence quenching of PTCBI by H₂Pc.²³ Thus, in Compartment B, the oxidizing power of the photogenerated holes can be consumed at the H₂Pc/water interface,¹³ leading to the oxidation of thiol (the formal potential of H₂Pc⁺/H₂Pc = +0.70 V vs. Ag/AgCl;¹³ the oxidation potential of thiol = +0.60 V vs. Ag/AgCl²⁴). In Compartment A (ITO/H₂Pc/C₆₀–Pt), the formation of the C₆₀²⁻ active species can occur under irradiation, which has been confirmed by means of in situ spectroelectrochemistry.¹⁴ Local accumulation of the C₆₀²⁻ species at the solid/water interface produces power for H⁺ reduction, resulting in the evolution of H₂ (the formal potential of H⁺/H₂ (pH = 2) = −0.32 V vs. Ag/AgCl). The potential of the electron-photodoped PTCBI in ITO/PTCBI/H₂Pc is more negative than that of the hole-photodoped H₂Pc in ITO/H₂Pc/C₆₀–Pt (i.e. the reduction potential of PTCBI/PTCBI⁻ = ∼0 V vs. Ag/AgCl¹³). Therefore, the hole-photodoped H₂Pc (H₂Pc⁺) can be converted back to H₂Pc by the electron transferred from the PTCBI layer via a connecting wire.

Conclusions

In summary, we have demonstrated a novel photocatalysis system in which the evolution of H₂ can be induced by organic p/n bilayers responsive to the entire spectrum of visible light. Some examples of two-step excitation photocatalysis systems (λ < ∼550 nm) have been reported in which both inorganic semiconductor photocatalysts were coupled with an electron relay system;^25–28 however, these are distinct from the present system in terms of the type of material, the structure of the photocatalysis system, and the range of visible light energy available to the photocatalytic reaction. Furthermore, considering the latter point, the present system is also distinct from the metal-free photocatalyst capable of H₂ production (photocatalysis response: λ < 550 nm) recently reported by Wang and Domen.²⁹ The efficiency of our photocatalyst system was compared with that of the reported visible- light photocatalyst systems for H₂ evolution in terms of EQE (Table S1 in the ESI†). These results also support the sugestion that the Pt-loaded H₂Pc/C₆₀ is a novel and advanced photocatalyst capable of H₂ evolution utilizing the entire visible light energy, because most of the reported EQE values were estimated in the near-UV region. Organic semiconductors offer easy processing and abundant variety for gaining better features of photocatalyst devices. As an example, the combination of an adsorbent (e.g. Nafion) with the PTCBI/H₂Pc bilayer was recently reported to afford efficient photocatalysis for the decomposition of a typical odorous compound, trimethylamine (ca. 40% of EQE for CO₂ formation).³⁰ Organophotocatalysts can be expected to open new avenues in the fields of photo/chemical energy conversion and large-scale photocatalysis, thus leading to innovative technology towards applications in interior living space as well as in the water phase.

Acknowledgements

This work was partly supported by a grant for Hirosaki University Institutional Research and a Grant-in-Aid for Scientific Research (T. A.) from the Ministry of Education, Culture, Sports, Science and Technology, Japan. We thank Dr Seiji Kakuta (Aomori Prefectural Industrial Technology Research Center) for the measurement of surface roughness.

References

1. A. Fujishima and K. Honda, Nature, 1972, 238, 37.
2. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503.
3. C.-C. Hu, Y.-L. Lee and H. Teng, J. Phys. Chem. C, 2011, 115, 2805.
4. J. Wang, B. Huang, Z. Wang, P. Wang, H. Cheng, Z. Zheng, X. Qin, X. Zhang, Y. Dai and M.-H. Whangbo, J. Mater. Chem., 2011, 21, 4562.
5. F. Zhang, K. Maeda, T. Takata and K. Domen, Chem. Commun., 2010, 46, 7313.
6. X. Qiu, M. Miyauchi, H. Yu, H. Irie and K. Hashimoto, J. Am. Chem. Soc., 2010, 132, 15259.
7. S. Ikeda, T. Nakamura, T. Harada and M. Matsumura, Phys. Chem. Chem. Phys., 2010, 12, 13943.
8. H. Kaga, K. Saito and A. Kudo, Chem. Commun., 2010, 46, 3779.
9. R. Abe, K. Shinmei, K. Hara and B. Ohtani, Chem. Commun., 2009, 3577.
10. T. Abe, K. Nagai, M. Kaneko, T. Okubo, K. Sekimoto, A. Tajiri and T. Norimatsu, ChemPhysChem, 2004, 5, 716.
11. T. Abe, K. Nagai, S. Kabutomori, M. Kaneko, A. Tajiri and T. Norimatsu, Angew. Chem., Int. Ed., 2006, 45, 2778.
12. T. Abe and K. Nagai, Org. Electron., 2007, 8, 262.
13. T. Abe, S. Miyakushi, K. Nagai and T. Norimatsu, Phys. Chem. Chem. Phys., 2008, 10, 1562.
14. T. Abe, S. Tobinai, N. Taira, J. Chiba, T. Itoh and K. Nagai, J. Phys. Chem. C, 2011, 115, 7701.
15. T. Shimidzu, T. Iyoda and Y. Koide, J. Am. Chem. Soc., 1985, 107, 35.
16. R. Abe, K. Sayama and H. Arakawa, Chem. Phys. Lett., 2003, 379, 230.
17. Q. Li, Z. Jin, Z. Peng, Y. Li, S. Li and G. Lu, J. Phys. Chem. C, 2007, 111, 8237.
18. T. Maki and H. Hashimoto, Bull. Chem. Soc. Jpn., 1952, 25, 411.
19. A. Capobianchi and M. Tucci, Thin Solid Films, 2004, 33, 451–452.
20. T. Morikawa, C. Adachi, T. Tsutsui and S. Saito, Nippon Kagaku Kaishi, 1990, 962–967.
21. J. H. Sharp and M. Lardon, J. Phys. Chem., 1968, 72, 3230.
22. K. Suemori, T. Miyata, M. Yokoyama and M. Hiramoto, Appl. Phys. Lett., 2004, 85, 6269–6271.
23. K. Nagai, Y. Fujimoto, H. Shiroishi, M. Kaneko, T. Norimatsu and T. Yamanaka, Chem. Lett., 2001, 354.
24. S. Surdhar and D. A. Armstrong, J. Phys. Chem., 1986, 90, 5915.
25. K. Tennakone and S. Wickramanayake, J. Chem. Soc., Faraday Trans. 2, 1986, 82, 1475.
26. K. Sayama, R. Yoshida, H. Kusama, K. Okabe, Y. Abe and H. Arakawa, Chem. Phys. Lett., 1997, 277, 387.
27. Y. Sasaki, H. Nemoto, K. Saito and A. Kudo, J. Phys. Chem. C, 2009, 113, 17536.
28. K. Maeda, R. Abe and K. Domen, J. Phys. Chem. C, 2011, 115, 3057.
29. X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8, 76.
30. K. Nagai, T. Abe, Y. Kaneyasu, Y. Yasuda, I. Kimishima, T. Iyoda and H. Imaya, ChemSusChem, 2011, 4, 727.

---

Footnote

† Electronic supplementary information (ESI) available: Absorption spectra of H₂Pc/C₆₀ and PTCBI/H₂Pc bilayers, and a list of efficiencies of visible light photocatalysts for H₂ evolution. See DOI: 10.1039/c2ra21281a

---