Efficient photocatalytic hydrogen evolution with end-group-functionalized cobaloxime catalysts in combination with graphite-like C₃N₄

Abstract

Three comparable hybrid photocatalytic systems, comprising semiconductor g-C₃N₄, end-group-functionalized cobaloxime complexes (carboxy-functionalized cobaloxime, C1; pyrene-functionalized cobaloxime, C2; and non-functionalized cobaloxime, C3), and triethanolamine (TEOA), are active for visible-light-driven hydrogen production in CH₃CN–H₂O (9/1, v/v) solution. Upon irradiation for 12 h, the turnover numbers of hydrogen evolution are 234, 281 and 195 for the hybrid systems C1/g-C₃N₄, C2/g-C₃N₄ and C3/g-C₃N₄, respectively. The highest hydrogen evolution efficiency of the C2/g-C₃N₄ system can be attributed to the strongest π–π interactions between the pyrene moiety and g-C₃N₄. Based on electrochemical properties, steady-state photoluminescence spectra and theoretical analyses, the visible light absorption of g-C₃N₄, the catalytic H₂-evolving ability of cobaloxime as well as the efficient charge separation of the excited g-C₃N₄ in the presence of both TEOA and cobaloxime, are responsible for the high activity of these hybrid systems.

---

Introduction

The visible-light-driven hydrogen evolution from water is one of the promising approaches to provide renewable and clean energy for the future.¹ Therefore, a number of hydrogen-evolving systems have been actively developed including homogeneous and heterogeneous photocatalytic systems in recent years.^1d,2 Studies on the homogeneous catalysts for hydrogen production are mainly focused on the first-row transition metal complexes, such as [FeFe]-hydrogenase mimics,³ cobalt and nickel complexes.⁴ Among various catalysts, the cobaloxime has been widely investigated for photocatalytic hydrogen evolution due to their convenient preparation and good catalytic activities.⁵ Many cobaloxime-based photocatalytic systems for hydrogen production have been constructed utilizing the noble-metal-based photosensitizers, such as ruthenium,⁶ rhenium,⁷ platinum,⁸ and iridium-based organometallic complexes.⁹ However, because of the high cost of noble-metal-based photosensitizers, chemists attempt to replace them with other chromophores made of earth-abundant elements, such as organic xanthene dyes,¹⁰ metalloporphyrins,¹¹ and photoactive nanomaterials.^2b,12

To our knowledge, Eisenberg and co-workers have employed the xanthene dye derivatives (S or Se in place of O in the xanthene ring) in combination with the cobaloxime for the light-driven generation of hydrogen, and reported the highest turnover of 9000 (vs. organic dye).^10b Besides the organic dyes, the metalloporphyrins were also introduced as the photosensitizers to couple with cobaloxime catalysts for efficient photoinduced hydrogen production.^11a,c Nevertheless, one of the inherent weaknesses of the utilized organic and organometallic photosensitizers is their instability upon long-term irradiation.¹³ Recently, an increasing number of noble-metal-free nanomaterials have been used as the light absorber, such as TiO₂,¹⁴ CdS,^13a core/shell CdSe/ZnS QDs,¹⁵ and graphite-like C₃N₄ (referred to hereafter as g-C₃N₄), which can link cobaloximes directly onto the surface and overcome the drawback of the poor stability of the organic and organometallic photosensitizers.¹⁶

Among various nanomaterials, the environmentally benign g-C₃N₄ is rightfully attracting increased attention owing to its relatively high stability and suitable electronic structure (E_g = 2.7 eV, conduction band at −1.42 V and valence band at 1.28 V vs. Ag/AgCl) covering the water-splitting potentials.¹⁷ However, g-C₃N₄ alone shows very poor photocatalytic activities for water reduction and relies on surface co-catalysts, which can accelerate the separation of photogenerated electron/hole and increase the photocatalytic performance of g-C₃N₄. The precedents for catalytic systems which combined g-C₃N₄ with a co-catalyst, such as platinum group metals and MoS₂,¹⁸ do exist but it is only recently that Co- and Ni-based molecular catalysts have been used as co-catalysts.^13b,16 Herein, three comparable cobaloxime complexes (C1, C2 and C3) (Fig. 1) are synthesized and studied as molecular co-catalysts for hydrogen generation from hybrid systems containing the photoactive g-C₃N₄ and triethanolamine (TEOA) as electron donor in acetonitrile aqueous solution. The results of photocatalysis experiments show that the TONs of hydrogen evolution are 234 and 281 (vs. complex) for the hybrid systems C1/g-C₃N₄ and C2/g-C₃N₄, respectively. In comparison, the photocatalytic activity of the non-functionalized complex C3 (TON = 195) is lower than that of the carboxy-functionalized complex C1 and pyrene-functionalized complex C2 under the same conditions. To better understand the difference of photocatalytic activities, the adsorption of the complexes to g-C₃N₄, the electrochemical properties and the steady-state photoluminescence properties are investigated and discussed in detail.

Fig. 1 Structures of g-C₃N₄ and complexes C1 to C3 used in the present work.

Result and discussion

Synthesis and structure of catalysts and g-C₃N₄

The g-C₃N₄ was prepared by the direct polymerization of urea at ambient conditions. The XRD pattern of the as-prepared g-C₃N₄ shows two characteristic peaks at 13.1 and 27.4 degree in accord with the previously reported results (Fig. S1†).¹⁹ The TEM image clearly shows the g-C₃N₄ possesses the layered structure (Fig. S2†). In order to investigate the effect of end-group-functionalized cobaloximes on the photocatalytic performance of as-prepared g-C₃N₄, three cobaloxime complexes are designed as co-catalysts and expected to assemble on the surface of g-C₃N₄ for photocatalytic H₂ evolution. The first one (C1) is based on the relatively good adsorption performance of carboxy group, which can provide a linkage to the nanomaterial surface.²⁰ Furthermore, in view of the conjugative π structure of g-C₃N₄,²¹ we choose another cobaloxime derivative C2, in which the pyrene moiety of π electron conjugation is expected to strongly interact with g-C₃N₄ via π–π interactions. As a comparison with the end-group-functionalized cobaloxime complexes C1 and C2, the complex C3 without functional-group modification is also employed as co-catalyst for g-C₃N₄ in this paper.

The C1 was conveniently prepared by stirring [Co(dmgH)(dmgH₂)Cl₂] with isonicotinic acid in CH₃CN and then obtained as dark-red crystals. The C2 was prepared in an analogous manner to C1 using [Co(dmgH)(dmgH₂)Cl₂] and N-pyren-1-ylmethyl-isonicotinamide, which was synthesized by the amidation reaction between isonicotinoyl chloride and (pyren-1-ylmethyl)amine. All new compounds were confirmed by ¹H, ¹³C NMR and elemental analysis. The solid-state structures of C1 and C2 were established by means of single-crystal X-ray diffraction. The crystal structures of complexes C1 and C2 are given in Fig. 2 with selected bond lengths and angles listed in Table 1. The coordination geometry around the Co-centre in C1 is strongly similar to those in other reported cobaloxime complexes.^8b Because of the existence of the carboxyl, the axial Co–N(5)_pyridine distance is 1.967 Å, slightly longer than that in C3 (1.959 Å). The average Co–N_imine bond distance of the glyoximate ligands in C1 is 1.896 Å, which is consistent to the value of 1.895 Å in C3.²² The asymmetric unit of C2 consists of two same Co-centred complexes and some uncoordinated solvent molecules. For clarity, we only present one Co-centre in Fig. 2, which is hexacoordinated with a slightly distorted octahedral geometry (Table 1). The pendent pyrene moiety is roughly parallel to the equatorial plane defined by the four N atoms of glyoximate ligands, with a dihedral angle of 5.6 degree.

Fig. 2 Crystal structures of C1 and C2 (the hydrogen atoms are omitted for clarity).

Table 1 Selected bond lengths (Å) and angles (degree) for C1 and C2

C1
Co(1)–N(1)	1.910(2)	N(3)–Co(1)–N(4)	98.64(10)
Co(1)–N(2)	1.887(2)	N(2)–Co(1)–N(4)	178.48(11)
Co(1)–N(3)	1.882(2)	N(3)–Co(1)–N(1)	177.92(11)
Co(1)–N(4)	1.906(2)	N(2)–Co(1)–N(1)	99.33(10)
Co(1)–N(5)	1.967(3)	N(4)–Co(1)–N(1)	80.26(11)
Co(1)–Cl(1)	2.2301(8)	N(3)–Co(1)–N(5)	89.63(10)
O(1)–N(1)	1.364(3)	N(2)–Co(1)–N(5)	89.41(10)
O(2)–N(2)	1.340(3)	N(4)–Co(1)–N(5)	92.07(10)
O(3)–N(3)	1.336(3)	N(1)–Co(1)–N(5)	92.17(10)
O(4)–N(4)	1.352(3)	N(3)–Co(1)–N(2)	81.72(10)
C2
Co(1)–N(1)	1.895(4)	N(3)–Co(1)–N(1)	177.37(18)
Co(1)–N(2)	1.904(4)	N(3)–Co(1)–N(4)	82.11(17)
Co(1)–N(3)	1.886(4)	N(1)–Co(1)–N(4)	98.27(17)
Co(1)–N(4)	1.898(4)	N(3)–Co(1)–N(2)	98.08(17)
Co(1)–N(9)	1.974(4)	N(1)–Co(1)–N(2)	81.48(17)
Co(1)–Cl(1)	2.2241(13)	N(4)–Co(1)–N(2)	178.84(17)
O(1)–N(1)	1.362(5)	N(3)–Co(1)–N(9)	90.92(17)
O(2)–N(2)	1.337(5)	N(1)–Co(1)–N(9)	91.68(17)
O(3)–N(3)	1.330(5)	N(4)–Co(1)–N(9)	90.07(15)
O(4)–N(4)	1.343(5)	N(2)–Co(1)–N(9)	91.07(16)

---

Table 2 Influence of the concentration of C1 and the amount of g-C₃N₄ on photocatalytic H₂ production^a

Run	Concentration of C1 (mM)	Amount of g-C3N4 (mg)	TON
a Condition: 5 vol% TEOA in CH3CN–H2O (9/1, v/v) at pH 10; irradiation time 4 h.
1	0.05	4	56
2	0.1	4	84
3	0.2	4	40
4	0.1	2	57
5	0.1	8	92

---

Adsorption of the cobaloximes C1–C3 to g-C₃N₄

The UV-vis diffuses reflectance absorption spectrum of g-C₃N₄ exhibits the broad absorption from UV to visible light region (Fig. S4†). The absorption edge is at about 455 nm, corresponding to a small band-gap (2.73 eV). The UV-vis absorption spectra of complexes C1–C3 measured in CH₃CN are shown in Fig. 3a. The cobaloximes C1 and C3 exhibit a high-energy absorption between 230 nm and 300 nm, which is attributed to the intra-ligand (π–π*) transitions.^8b Unlike C1 and C3, the C2 displays a very different absorption spectrum with characteristic absorption bands of the pyrene group (at 342, 326, 312, 276, 265, 255, 242, and 235 nm).²³ According to the report by Li et al., the cobaloximes bearing appropriate pendent ligands can be effectively adsorbed by the photoactive nanomaterials.^13a To investigate the adsorption amount of the cobaloximes C1–C3 on the g-C₃N₄ surface, the acetonitrile solutions of C1–C3 (2 × 10⁻⁵ M, 5 mL) were measured by UV-vis spectrophotometry before and after exposure to g-C₃N₄ (Fig. 3). In view of the absorbance difference at the maximum adsorption wavelength, we estimate the adsorption amount of C1, C2, and C3 on g-C₃N₄ to be approximately 0.05, 0.07, and 0.02 μmol, respectively. Increasing the original concentration of cobaloximes from 2 × 10⁻⁵ M to 3 × 10⁻⁵ M did not lead to the increase in adsorption amount. As noted, however, no significant difference in the adsorption amount was seen between the above-mentioned cobaloximes, with a median adsorption amount for each of about 0.04 μmol.

Fig. 3 (a) UV-vis absorption spectra of C1 (2 × 10⁻⁵ M), C2 (2 × 10⁻⁵ M) and C3 (2 × 10⁻⁵ M) in CH₃CN (5 mL); UV-vis absorption spectra of C2 (b), C1 (c), and C3 (d) before and after stirred with g-C₃N₄ (4 mg) in CH₃CN for 2 h following centrifugation and filtration.

Photocatalytic activities of the hybrid systems

We first studied the photocatalytic H₂ production of the hybrid system C1/g-C₃N₄ in CH₃CN–H₂O (9 : 1, v/v, 5 mL) solution containing 5 vol% TEOA at pH 10. Experiments under irradiation (λ > 400 nm) with different concentration of C1 (0.05–0.2 mM) showed the system with 0.1 mM C1 released the largest amount of H₂ (42 μmol, TON = 84) over 4 h (Table 2). With an increase or a decrease of the concentration of C1 to 0.2 mM and 0.05 mM, the TON of H₂ evolution decreased to 40 and 56, respectively. It is evident that the optimum amount (0.5 μmol) of C1 for the system is much higher than the adsorption amount (0.05 μmol) of C1 on the surface of g-C₃N₄, which indicates that the complex C1 predominantly exists as free molecules in reaction solution. To further clarify the role of C1, the suspension of the C1/g-C₃N₄ system was stirred for 2 h in the dark and subsequently filtered, resulting in the g-C₃N₄ residue and the clear filtrate. With readdition of reaction solution, the g-C₃N₄ residue adsorbing a small amount of C1 (0.05 μmol) gave a quarter of the TON of the original C1/g-C₃N₄ system, while with readdition of fresh g-C₃N₄, the clear filtrate containing the free molecules displayed a H₂ evolution rate similar to that of the original system C1/g-C₃N₄ (Fig. S6†). These results indicate that both of the C1 molecules, adsorbed on the surface of g-C₃N₄ and dissolved in reaction solution, can improve the photocatalytic activities of g-C₃N₄. The light-induced H₂ production catalyzed by C1 also depends on the amount of g-C₃N₄. When the concentration of C1 was 0.1 mM, the amount of H₂ evolved was improved apparently with an increase of the amount of g-C₃N₄ from 2 mg to 4 mg. However, further increasing the amount of g-C₃N₄ to 8 mg did not lead to an obvious enhancement in the photocatalytic activity. Control experiments without C1, g-C₃N₄, or TEOA showed no H₂ was released from the system C1/g-C₃N₄, suggesting that all three components are required for the photocatalytic H₂ evolution.

Besides CH₃CN, CH₃CH₂OH and DMF were also used as reaction solvent for the photocatalytic experiment in the C1/g-C₃N₄ system. Smaller quantities of hydrogen were released in systems containing either CH₃CH₂OH or DMF, albeit with the same 9/1 v/v ratio with water (Fig. 4). Consequently, the solvent CH₃CN is superior to CH₃CH₂OH and DMF for H₂ evolution in this system. Further studies on solvent effects were done to investigate the photocatalytic activity of the system C1/g-C₃N₄ by changing the ratio of CH₃CN and H₂O. The results showed that the C1/g-C₃N₄ system displayed a higher activity when the CH₃CN–H₂O ratio was increased. With an increase of CH₃CN–H₂O ratio to 9/1 from 1/1, the TON of H₂ evolution was apparently increased to 88 from 10. In addition to the system C1/g-C₃N₄, the effects of the two factors (i.e., organic solvents and CH₃CN–H₂O ratios) on the photocatalytic performances in the analogous systems C2/g-C₃N₄ and C3/g-C₃N₄ were also studied in Fig. 4. It is found that the trends of the medium dependence on hydrogen evolution over the C2/g-C₃N₄ and C3/g-C₃N₄ systems agree closely with those over the C1/g-C₃N₄ system (Fig. 4).

Fig. 4 Effects of organic solvents and CH₃CN–H₂O ratios on the hydrogen evolution over each system containing cobaloxime (1 × 10⁻⁴ M), g-C₃N₄ (4 mg) and TEOA (5 vol%) at pH 10; irradiation time 6 h.

Considering the aforementioned catalytic condition, the pH value of the C1/g-C₃N₄ system is approximately 10 without any adjustment by acid or base. To examine the pH effects on H₂ evolution from the system C1/g-C₃N₄, light-induced H₂ evolution was performed in the pH range of 8.5–10 under the conditions described in Fig. 5. The results show that the photocatalytic H₂ production is dependent on the pH value of the system, in which the optimum pH value for H₂ evolution is 9, while lower amounts of H₂ are obtained at either lower or higher pH values. The mechanism on the similar pH-dependent phenomenon has been elucidated in more detail by Zhang et al.^10c In addition, the pH dependence on hydrogen production over the systems C2/g-C₃N₄ and C3/g-C₃N₄ also shows that the maximum hydrogen generation efficiency was achieved at pH 9. This pH-dependent effect is related to the oxidation of TEOA, which is an essential step in the catalytic cycle.^8a As we know, TEOA is extensively used as electron donor in many Pt/g-C₃N₄-based hydrogen evolution systems.^19a,21a,24 In the C1/g-C₃N₄ system, keeping the concentration of C1 (1 × 10⁻⁴ M) and the amount of g-C₃N₄ (4 mg) unchanged, increasing the concentration of TEOA from 2.5 to 5 vol% apparently improved the efficiency of H₂ production. Further increasing the concentration of TEOA to 10 vol%, however, led to no obvious increase in the amount of H₂ production (Fig. S9†). As a result, an optimized H₂ evolution system containing 1 × 10⁻⁴ M C1, 4 mg g-C₃N₄, and 5 vol% TEOA in 5 mL CH₃CN–H₂O (9/1, v/v) at pH 9.0 was able to photocatalyze H₂ evolution with a TON of 234.

Fig. 5 pH dependence on the hydrogen production over each system under the following conditions: TEOA (5 vol%), cobaloxime (1 × 10⁻⁴ M) and g-C₃N₄ (4 mg) in CH₃CN–H₂O (9/1, v/v) solution; irradiation time 12 h.

To examine the effect of carboxyl group on the photocatalytic activity, the photocatalytic H₂ production experiment by the C3/g-C₃N₄ system was studied under the same conditions. A total amount of evolved H₂ reached 97 μmol, corresponding to a TON of 195. The activity of the C1/g-C₃N₄ system is higher than that of the C3/g-C₃N₄ system, indicating that the carboxy group functions satisfactorily as a linkage for H₂ evolution. Furthermore, it was found that the C2/g-C₃N₄ system displayed the highest activity among these three complexes, with TONs up to 281(Fig. 6). As expected, both of the C2 molecules, adsorbed on g-C₃N₄ and dissolved in solution, also displayed the highest H₂ evolution efficiency among three (Fig. S7†). These results indicate that the C2 is the best co-catalyst for H₂ production in the current photocatalytic system, presumably due to π–π interactions between C2 and g-C₃N₄.

Fig. 6 Time dependence of hydrogen production using different cobaloxime complexes C1 to C3 (1 × 10⁻⁴ M) under the following conditions: g-C₃N₄ (4 mg) and TEOA (5 vol%) in CH₃CN–H₂O (9/1, v/v) solution (5 mL) at pH 9, error bars represent standard errors of the means of three independent experiments.

The photocatalytic H₂ production of the C1/g-C₃N₄ system leveled out after 9 h of irradiation (Fig. 6). To investigate the reason(s) for the cease of H₂ generation, after the photocatalysis reaction, the suspension was filtered and the separated solid was washed with water and dried in vacuum. The UV-vis absorption spectrum of the resulting clear solution is different from that of the original C1 (Fig. S5†). Thus we speculate that the cessation of H₂ production may be ascribed to the degradation of C1. Besides the degradation of C1 after photocatalysis reaction, the pH value of system changed to 9.6 from 9 after photocatalytic process. After 12 h of irradiation, to recover H₂ evolution of the C1/g-C₃N₄ system, the photocatalytic experiment with readdition of C1 (0.5 μmol) to the system showed a recovered but lowered activity for hydrogen evolution under irradiation (Fig. S10†). For comparison, CoCl₂ was also tested for the photocatalytic H₂ evolution with g-C₃N₄ under the same conditions. The results show that the photocatalytic activity of the CoCl₂/g-C₃N₄ system is fairly lower than that of the C1/g-C₃N₄ system (Fig. S11†), suggesting that the active species in the current system is the cobaloxime C1 rather than the decomposed products. The TEM image of the separated solid is as shown in Fig. S3,† which is similar to the TEM image before irradiation, indicating that g-C₃N₄ is stable in the photocatalytic process. No zero-valent cobalt colloids were observed in the TEM image, which further proved the active species is also not the zero-valent cobalt in this system. The XRD pattern of the isolated solid is very similar to that of the freshly prepared g-C₃N₄, except for a decrease in the intensity of the peak at 27.4 degree (Fig. S1†).

Electrochemical properties

The redox potential of the complexes is crucial to understand the electron transfer in photocatalytic process. Cyclic voltammetry studies were performed in CH₃CN with Bu₄NPF₆ as supporting electrolyte under N₂. The electrochemical reduction processes of C1–C3 are shown in Fig. S12.† The reduction potentials are versus Fc⁺/Fc and are summarized in Table 3. All Co^III complexes have two reductions, which can be assigned to Co^III/Co^II and Co^II/Co^I, respectively.²⁵ The first irreversible reduction potentials of C1, C2 and C3 are confirmed to be at about −1.06, −1.13, and −1.22 V, respectively. The first potentials of C1 and C2 are anodic shift relative to C3, which may be attributable to the axial pyridine ligand modified by electron-withdrawing group (–COOH or –CONH). But there is no above trend in the second potentials for Co^II/Co^I couple. The potentials at the secondary reduction are nearly equal for the three complexes. The behaviour of electrocatalytic proton reduction by C1 or C2 was studied by cyclic voltammograms in the presence of CH₃COOH (Fig. S13 and S14†). Overall, complexes C1 and C2 behaved analogously toward electrocatalytic proton reduction.

Table 3 Electrochemical potentials^a (vs. Fc⁺/Fc) for all cobaloximes

Cobaloximes	Co^III/Co^II E1/2^b	Co^II/Co^I E1/2	ΔG^c
a Electrochemical potentials were obtained by cyclic voltammetry studies under a N2 atmosphere with 0.1 M Bu4NPF6 as the supporting electrolyte in CH3CN.b Irreversible reduction wave.c ΔG = the potential of CB(g-C3N4) − E1/2(Co^II/Co^I), CB(g-C3N4) = −1.87 V.
C1	−1.06	−1.56	−0.31
C2	−1.13	−1.48	−0.39
C3	−1.22	−1.50	−0.37

---

The probable mechanism on the photocatalytic H₂ production

The estimated reduction potential of g-C₃N₄ is −1.42 V relative to Ag/AgCl,^18a which can be adjusted to −1.87 V versus Fc⁺/Fc (E⁰ of Fc⁺/Fc = 0.45 V vs. Ag/AgCl).²⁶ The reduction potentials of the catalysts C1–C3 are given in Table 3. According to Weller-equation, the values of free-energy changes (ΔG) for formation of Co^I species are determined to be −0.31, −0.39, and −0.37 V for each system, respectively (Table 3). The negative ΔG suggests that it is thermodynamically feasible for photoinduced electron transfer from the conduction band of g-C₃N₄ to these catalysts. However, there is no significant difference in the ΔG value between the three hybrid systems.

To explore the charge transfer pathways, the steady-state photoluminescence properties of the C2/g-C₃N₄ system were investigated in acetonitrile aqueous solution. The g-C₃N₄ suspension excited at 375 nm results in a maximal fluorescence emission peak at 459 nm (Fig. 7a). No significant decreases in fluorescence intensity are observed upon addition of either C2 or TEOA. The reason for this phenomenon may plausibly be attributed to the fast recombination of photogenerated electron/hole pairs. However, a larger fluorescence quenching of g-C₃N₄ by 25% upon addition of both TEOA and C2 is observed, suggesting that both C2 and TEOA are required for the separation of the photogenerated electron/hole. Based on hydrogen production experiment and the above spectroscopic study, a plausible pathway of the photocatalytic H₂ evolution reaction is proposed in Fig. 8. Under visible light irradiation, the excited state electrons of the valence band of g-C₃N₄ would transport to the conduction band. The conduction-band electron of g-C₃N₄ can give two electrons to C2 to form to Co^I species. The photogenerated Co^I species is protonated to obtain a Co–H intermediate, which finally reacts with another proton to generate H₂.^5a Meanwhile the electron donor TEOA is oxidized by the photogenerated holes in the valence band.

Fig. 7 (a) The emission spectra of g-C₃N₄ (0.8 g L⁻¹, 5 mL) suspension in CH₃CN–H₂O (9/1, v/v) solution at pH 9 in the presence of C2 (1 × 10⁻⁴ M), TEOA (5 vol%), TEOA (5 vol%) and C2 (1 × 10⁻⁴ M); (b) the fluorescence quenching of g-C₃N₄ in the presence of C1 (1 × 10⁻⁴ M), C2 (1 × 10⁻⁴ M), and C3 (1 × 10⁻⁴ M) in 5 vol% TEOA acetonitrile aqueous solution.

Fig. 8 A proposed pathway for the photocatalytic H₂ evolution by cobaloxime–g-C₃N₄ systems.

In contrast, the fluorescence intensity of g-C₃N₄ decreased by 20% and 14% upon addition of C1 + TEOA and C3 + TEOA (Fig. 7b), respectively. The sequence of the effectiveness quenching fluorescence of g-C₃N₄ is C2 > C1 > C3, which corresponds to the order of their photocatalytic H₂ evolution activity. According to previous reports on hybrid photocatalytic systems containing cobaloxime, the photocatalytic activity could be influenced by the adsorption amount of cobaloximes on nanomaterials, the driving force (ΔG) of electron transfer and the attachment mode between cobaloxime and nanomaterials.^{13a,14a,15,16} However, no apparent difference in terms of the adsorption amount and the driving force was found between the present three systems. Therefore, the reason for the activity difference may be attributed to the attachment mode between g-C₃N₄ and cobaloximes. Compared with the free-collision mode between the non-functionalized C3 and g-C₃N₄, the direct attachment (i.e., carboxy linkage or π–π interactions) between end-group functionalized cobaloxime and g-C₃N₄ could promote the charge transfer from g-C₃N₄ to the co-catalytic centres.

Conclusions

Three end-group-functionalized cobaloxime complexes as co-catalyst were designed and assembled on the surface of g-C₃N₄ for photocatalytic H₂ evolution in CH₃CN–H₂O (9/1, v/v) solution at pH 9. The photocatalytic experiments gave the highest TON of 234, 281 and 195 for C1/g-C₃N₄, C2/g-C₃N₄ and C3/g-C₃N₄ hybrid systems, respectively. The photocatalytic activity of these systems was attributed to the efficient charge separation of the excited g-C₃N₄, which was proved by steady-state photoluminescence spectra. Among the cobaloximes, the C2 was superior to C1 and C3 as co-catalyst for H₂ production, which may result from the π–π interactions between C2 and g-C₃N₄. Due to the degradation of cobaloxime, the catalytic lifetime of the hybrid system is short, which motivates us to seek more stable and effective molecule co-catalysts for g-C₃N₄.

Experimental section

Chemicals

All chemicals in this work were purchased from commercial sources and used without further purification. (Pyren-1-ylmethyl)amine,²⁷ complexes [Co(dmgH) (dmgH₂)Cl₂],²⁸ and [Co(dmgH)₂pyCl] (C3)²⁸ were synthesized according to previously described procedures.

Synthesis

The preparation of g-C₃N₄ was following a previously reported method.¹⁹ Briefly, 10 g dry urea was added to an alumina crucible with a cover under ambient pressure in air. The precursor was put in a muffle furnace and heated to 550 °C in 2 h and maintained at this temperature for 3 h. The obtained powders were rinsed several times with distilled water and ethanol and dried in vacuum at 80 °C for 8 h.

N-Pyren-1-ylmethyl-isonicotinamide. A mixture of isonicotinic acid (0.30 g, 2.44 mmol) and thionyl chloride (20 mL) in a round-bottomed flask was refluxed for 5 h with stirring. Excess thionyl chloride was removed under reduced pressure until yellow powders of isonicotinoyl chloride were obtained. A solution of the prepared isonicotinoyl chloride in 20 mL of CH₂Cl₂ was treated with a mixed solution of (pyren-1-ylmethyl)amine (0.46 g, 2.00 mmol) and triethylamine (3 mL) in 20 mL of CH₂Cl₂ with vigorous stirring at 0 °C. After addition, the resulting mixture was allowed to react for another 8 h at room temperature. Finally, the solvent was evaporated, and the residue was purified by chromatography on a silica gel column. Elution with CH₃OH–CH₂Cl₂ (1 : 5, v/v) furnished pure N-pyren-1-ylmethyl-isonicotinamide (0.61 g, 91%). Anal. calcd for C₂₃H₁₆N₂O: C, 82.12%; H, 4.79%; N, 8.33%. Found: C, 82.09%; H, 4.81%; N, 8.29%. ¹H NMR (400 MHz, CDCl₃) δ 8.64 (2H, d, J = 5.4 Hz), 8.27 (1H, d, J = 9.2 Hz), 8.24–8.19 (2H, m), 8.16 (2H, d, J = 8.6 Hz), 8.05 (4H, dt, J = 16.9 Hz, 8.7 Hz), 7.58 (2H, d, J= 5.6 Hz), 6.63 (1H, s), 5.34 (2H, d, J = 5.1 Hz). ¹³C NMR (101 MHz, CDCl₃) δ 165.06, 150.28, 141.50, 131.55, 131.23, 130.68, 130.06, 129.19, 128.64, 127.85, 127.49, 127.31, 126.28, 125.67, 125.56, 125.11, 124.86, 124.66, 122.51, 120.97, 42.80.

[Co(dmgH)₂(4-COOH-py)Cl] (C1). A 0.36 g (1.00 mmol) sample of [Co(dmgH)(dmgH₂)Cl₂] was suspended in 40 mL of CH₃CN. One equivalent of isonicotinic acid (0.12 g, 1.00 mmol) was then added to the flask and heated at 70 °C with stirring for 1 h. The resulting dark-red solution was filtered and left overnight. Dark-red crystals (0.39 g, 87%) of suitable for X-ray analysis were obtained. Anal. calcd for CoClC₁₄N₅O₆H₁₉·H₂O: C, 36.10%; H, 4.54%; N, 15.04%. Found: C, 36.47%; H, 4.44%; N, 15. 16%. ¹H NMR (400 MHz, CD₃CN) δ 8.30 (2H, d, J = 6.8 Hz), 7.72 (2H, d, J = 6.7 Hz), 2.31 (12H, s).

[Co(dmgH)₂(N-pyren-1-ylmethyl-isonicotinamide)Cl] (C2). Triethylamine (0.10 g, 1.00 mmol) was added to a stirred, green suspension of [CoCl₂(dmgH)(dmgH2)] (0.36 g, 1.00 mmol) in CH₃OH (25 mL), resulting in a brown solution. A solution of N-pyren-1-ylmethyl-isonicotinamide (0.33 g, 1.00 mmol) in 15 mL of CH₂Cl₂ was added after five minutes and the reaction mixture was heated at 40 °C for 1 h. The resulting brown solution was filtered and left overnight. Brown crystals (0.50 g, 76%) of suitable for X-ray analysis were obtained. Anal. calcd for CoClC₃₁N₆O₅H₃₀: C, 56.33%; H, 4.57%; N, 12.71%. Found: C, 56.47%; H, 4.62%; N, 12.60%. ¹H NMR (400 MHz, CDCl₃) δ 8.32 (2H, d, J = 4.8 Hz), 8.22 (3H, d, J = 7.8 Hz), 8.14 (2H, d, J = 9.2 Hz), 8.11–8.00 (3H, m), 7.98 (1H, d, J = 7.4 Hz), 7.55 (2H, d, J = 4.8 Hz), 6.71 (1H, s), 5.30 (2H, s), 2.31 (12H, s).

Adsorption experiments: a solution of C1, C2 or C3 (2 × 10⁻⁵ M, 5 mL) in CH₃CN solution was added to g-C₃N₄ (4 mg) in a Schlenk tube. The mixture was stirred for 2 h in the dark and subsequently centrifuged. The filtered, clear solution was then measured by UV-visible absorption spectra. The amount of C1, C2 or C3 adsorbed on g-C₃N₄ was estimated by the absorbance difference at the absorption peak before and after adsorption on to g-C₃N₄.

Physical measurements

Elemental analyses were carried out on a Vario MICRO Elemental Analyser. ¹H, ¹³C NMR spectra were performed on a Bruker Avance III (400 MHz) spectrometer. TEM analyses were conducted on a JEM-2010 electron microscope at an acceleration voltage of 200 kV. UV-vis absorption spectra were measured on a Perkin-Elmer Lambda 35 UV-vis spectrophotometer.

X-ray diffraction

X-ray single-crystal data of C1 and C2 were collected on a SuperNova, Dual, Cu at zero, Atlas diffractometer. Crystal data collection, refinement and reduction were accomplished with the CrysAlisPro, Agilent Technologies, Version 1.171.36.28. The crystal structures were solved by direct methods with SHELXS-97 and refined by using the SHELXL-97 crystallographic software package. All non-hydrogen atoms were refined anisotropically. The hydrogen atoms were added in a riding model. Details of crystal data are summarized in Table 4. X-ray diffraction (XRD) was used to identify the structure of the g-C₃N₄ before and after photocatalytic reactions. Diffraction date were collected on a MiniFlex II diffractometer with Cu Kα radiation (30 kV × 15 mA). The 2θ scanning range was from 10 degree to 50 degree with a scanning speed of 4 degree per min.

Table 4 Crystal data and structure refinement details for complex C1 and C2

Complex	C1·(H2O)(CH3CN)	C2·(CH3OH)(H2O)2(CH2Cl2)
Molecule formula	C16H24ClCoN6O7	C64H70Cl4Co2N12O13
Formula weight	506.79	1474.98
T (K)	100.00(13)	100.0(2)
Crystal system	Triclinic	Triclinic
Space group	P1̄	P1̄
a (Å)	7.9393(3)	9.5141(4)
b (Å)	11.3829(4)	17.4267(9)
c (Å)	13.7720(6)	20.3947(10)
α (°)	111.729(4)	75.820(5)
β (°)	101.772(3)	87.861(4)
γ (°)	99.800(3)	74.763(4)
V (Å^3)	1089.74(7)	3161.8(3)
Z	2	2
F (000)	524	1528
Dcalc (g cm^−3)	1.544	1.549
R1/wR2 (I > 2σ(I))	0.0394/0.0860	0.0715/0.1822
R1/wR2 (all data)	0.0447/0.0883	0.0994/0.2066
Goodness of fit	1.084	1.032

---

Electrochemistry

Electrochemical measurements were made using a CH instrument Model 630A Electrochemical Workstation. The cyclic voltammetry experiments were conducted in acetonitrile solution containing 0.1 M ⁿBu₄NPF₆ as the supporting electrolyte under N₂. Glassy carbon and platinum wire were used as the working and counter electrodes, respectively, and the potential was measured against the Ag/AgCl reference electrode and reported relative to the internal reference of Fc⁺/Fc = 0.00 V.

Photocatalysis

In a typical procedure, g-C₃N₄ (4 mg), C1 (500 μL, 1 mM) and 5 vol% triethanolamine (TEOA) acetonitrile aqueous (9/1, v/v) solution were added to a Schlenk bottle. The mixture was magnetically stirred under N₂ atmosphere for 15 min. The system was freeze–pump–thaw degassed for three times and then warmed to room temperature prior to irradiation. The reaction solution was irradiated at 25 °C using a Xe lamp (300 W) with a cutoff filter (λ > 400 nm). The gas phase of the reaction system was analyzed on a GC 7900 instrument with a 5 Å molecular sieve column, a thermal conductivity detector, and using N₂ as carrying gas. The amount of hydrogen generated was determined by the external standard method. Hydrogen dissolved in the solution was not measured and the slight effect of the hydrogen generated on the pressure of the Schlenk bottle was neglected for calculation of the volume of hydrogen gas.

Acknowledgements

The authors thank the National Natural Science Foundation China (grant numbers 21071145, 21101153, 21231003 and 21203195) for financial support.

Notes and references

1. (a) N. S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 15729–15735; (b) R. M. Navarro, M. C. Alvarez-Galvan, J. A. Villoria de la Mano, S. M. Al-Zahrani and J. L. G. Fierro, Energy Environ. Sci., 2010, 3, 1865–1882; (c) E. S. Andreiadis, M. Chavarot-Kerlidou, M. Fontecave and V. Artero, Photochem. Photobiol., 2011, 87, 946–964; (d) W. T. Eckenhoff and R. Eisenberg, Dalton Trans., 2012, 41, 13004–13021.
2. (a) A. Kudo, Int. J. Hydrogen Energy, 2007, 32, 2673–2678; (b) P. D. Tran, L. H. Wong, J. Barber and J. S. C. Loo, Energy Environ. Sci., 2012, 5, 5902–5918; (c) V. Artero and M. Fontecave, Chem. Soc. Rev., 2013, 42, 2338–2356.
3. (a) M. Wang, L. Chen, X. Li and L. Sun, Dalton Trans., 2011, 40, 12793–12800; (b) F. Wang, W. G. Wang, H. Y. Wang, G. Si, C. H. Tung and L. Z. Wu, ACS Catal., 2012, 2, 407–416.
4. (a) S. Losse, J. G. Vos and S. Rau, Coord. Chem. Rev., 2010, 254, 2492–2504; (b) W. Zhang, J. Hong, J. Zheng, Z. Huang, J. Zhou and R. Xu, J. Am. Chem. Soc., 2011, 133, 20680–20683; (c) Z. Han, W. R. McNamara, M. S. Eum, P. L. Holland and R. Eisenberg, Angew. Chem., Int. Ed., 2012, 51, 1667–1670; (d) H. H. Cui, J. Y. Wang, M. Q. Hu, C. B. Ma, H. M. Wen, X. W. Song and C. N. Chen, Dalton Trans., 2013, 42, 8684–8691; (e) H. N. Kagalwala, E. Gottlieb, G. Li, T. Li, R. Jin and S. Bernhard, Inorg. Chem., 2013, 52, 9094–9101.
5. (a) V. Artero, M. Chavarot-Kerlidou and M. Fontecave, Angew. Chem., Int. Ed., 2011, 50, 7238–7266; (b) M. Wang, L. Chen and L. Sun, Energy Environ. Sci., 2012, 5, 6763–6778.
6. A. Fihri, V. Artero, M. Razavet, C. Baffert, W. Leibl and M. Fontecave, Angew. Chem., Int. Ed., 2008, 120, 574–577.
7. (a) B. Probst, M. Guttentag, A. Rodenberg, P. Hamm and R. Alberto, Inorg. Chem., 2011, 50, 3404–3412; (b) M. Guttentag, A. Rodenberg, R. Kopelent, B. Probst, C. Buchwalder, M. Brandstätter, P. Hamm and R. Alberto, Eur. J. Inorg. Chem., 2012, 2012, 59–64.
8. (a) P. Du, K. Knowles and R. Eisenberg, J. Am. Chem. Soc., 2008, 130, 12576–12577; (b) P. Du, J. Schneider, G. Luo, W. W. Brennessel and R. Eisenberg, Inorg. Chem., 2009, 48, 4952–4962.
9. A. Fihri, V. Artero, A. Pereira and M. Fontecave, Dalton Trans., 2008, 5567–5569.
10. (a) T. Lazarides, T. McCormick, P. Du, G. Luo, B. Lindley and R. Eisenberg, J. Am. Chem. Soc., 2009, 131, 9192–9194; (b) T. M. McCormick, B. D. Calitree, A. Orchard, N. D. Kraut, F. V. Bright, M. R. Detty and R. Eisenberg, J. Am. Chem. Soc., 2010, 132, 15480–15483; (c) P. Zhang, M. Wang, J. Dong, X. Li, F. Wang, L. Wu and L. Sun, J. Phys. Chem. C, 2010, 114, 15868–15874; (d) J. Dong, M. Wang, P. Zhang, S. Yang, J. Liu, X. Li and L. Sun, J. Phys. Chem. C, 2011, 115, 15089–15096; (e) L. Gong, J. Wang, H. Li, L. Wang, J. Zhao and Z. Zhu, Catal. Commun., 2011, 12, 1099–1103; (f) T. M. McCormick, Z. Han, D. J. Weinberg, W. W. Brennessel, P. L. Holland and R. Eisenberg, Inorg. Chem., 2011, 50, 10660–10666.
11. (a) P. Zhang, M. Wang, C. Li, X. Li, J. Dong and L. Sun, Chem. Commun., 2010, 46, 8806–8808; (b) K. Peuntinger, T. Lazarides, D. Dafnomili, G. Charalambidis, G. Landrou, A. Kahnt, R. P. Sabatini, D. W. McCamant, D. T. Gryko, A. G. Coutsolelos and D. M. Guldi, J. Phys. Chem. C, 2012, 117, 1647–1655; (c) M. Natali, R. Argazzi, C. Chiorboli, E. Iengo and F. Scandola, Chem.–Eur. J., 2013, 19, 9261–9271.
12. F. Wen and C. Li, Acc. Chem. Res., 2013, 46, 2355–2364.
13. (a) F. Wen, J. Yang, X. Zong, B. Ma, D. Wang and C. Li, J. Catal., 2011, 281, 318–324; (b) J. Dong, M. Wang, X. Li, L. Chen, Y. He and L. Sun, ChemSusChem, 2012, 5, 2133–2138.
14. (a) F. Lakadamyali and E. Reisner, Chem. Commun., 2011, 47, 1695–1697; (b) F. Lakadamyali, M. Kato, N. M. Muresan and E. Reisner, Angew. Chem., Int. Ed., 2012, 51, 9381–9384; (c) F. Lakadamyali, A. Reynal, M. Kato, J. R. Durrant and E. Reisner, Chem.–Eur. J., 2012, 18, 15464–15475.
15. J. Huang, K. L. Mulfort, P. Du and L. X. Chen, J. Am. Chem. Soc., 2012, 134, 16472–16475.
16. S. W. Cao, X. F. Liu, Y. P. Yuan, Z. Y. Zhang, J. Fang, S. C. J. Loo, J. Barber, T. C. Sum and C. Xue, Phys. Chem. Chem. Phys., 2013, 15, 18363–18366.
17. X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8, 76–80.
18. (a) J. Zhang, X. Chen, K. Takanabe, K. Maeda, K. Domen, J. D. Epping, X. Fu, M. Antonietti and X. Wang, Angew. Chem., Int. Ed., 2010, 49, 441–444; (b) X. H. Li, X. Wang and M. Antonietti, Chem. Sci., 2012, 3, 2170–2174; (c) Y. Hou, A. B. Laursen, J. Zhang, G. Zhang, Y. Zhu, X. Wang, S. Dahl and I. Chorkendorff, Angew. Chem., Int. Ed., 2013, 52, 3621–3625.
19. (a) S. Min and G. Lu, J. Phys. Chem. C, 2012, 116, 19644–19652; (b) Y. Zhang, J. Liu, G. Wu and W. Chen, Nanoscale, 2012, 4, 5300–5303.
20. (a) M. Sykora, M. A. Petruska, J. Alstrum-Acevedo, I. Bezel, T. J. Meyer and V. I. Klimov, J. Am. Chem. Soc., 2006, 128, 9984–9985; (b) J. Huang, D. Stockwell, Z. Huang, D. L. Mohler and T. Lian, J. Am. Chem. Soc., 2008, 130, 5632–5633; (c) A. J. Morris-Cohen, M. T. Frederick, L. C. Cass and E. A. Weiss, J. Am. Chem. Soc., 2011, 133, 10146–10154.
21. (a) K. Takanabe, K. Kamata, X. Wang, M. Antonietti, J. Kubota and K. Domen, Phys. Chem. Chem. Phys., 2010, 12, 13020–13025; (b) Y. Wang, Z. Wang, S. Muhammad and J. He, CrystEngComm, 2012, 14, 5065–5070.
22. S. Geremia, R. Dreos, L. Randaccio, G. Tauzher and L. Antolini, Inorg. Chim. Acta, 1994, 216, 125–129.
23. V. Georgakilas, K. Kordatos, M. Prato, D. M. Guldi, M. Holzinger and A. Hirsch, J. Am. Chem. Soc., 2002, 124, 760–761.
24. (a) X. Wang, K. Maeda, X. Chen, K. Takanabe, K. Domen, Y. Hou, X. Fu and M. Antonietti, J. Am. Chem. Soc., 2009, 131, 1680–1681; (b) Y. Zheng, J. Liu, J. Liang, M. Jaroniec and S. Z. Qiao, Energy Environ. Sci., 2012, 5, 6717–6731.
25. M. Razavet, V. Artero and M. Fontecave, Inorg. Chem., 2005, 44, 4786–4795.
26. G. A. N. Felton, C. A. Mebi, B. J. Petro, A. K. Vannucci, D. H. Evans, R. S. Glass and D. L. Lichtenberger, J. Organomet. Chem., 2009, 694, 2681–2699.
27. H. Lee, E. Luna, M. Hinz, J. J. Stezowski, A. S. Kiselyo and R. G. Harvey, J. Org. Chem., 1995, 60, 5604–5613.
28. W. C. Trogler, R. C. Stewart, L. A. Epps and L. G. Marzilli, Inorg. Chem., 1974, 13, 1564–1570.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1–S14. CCDC 955344 and 982938. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4ra01413h

---