In order to meet long-term global human energy demands,a sustainable alternative to fossil fuels is conversion of solar energy into a fuel which carries high energetic density stored within chemical bonds,such as hydrogen ^[1]. Artificial photosynthesis splitting water into hydrogen and oxygen is one of the most promising strategies for sustainable energy cycles ^{[2, 3]},because its oxidation ‘back’ to water in a fuel-cell efficiently restitutes the stored energy,in the form of electricity and without any chemical waste ^{[4, 5]}. To that aim,photoelectrochemical cells (PECs) have been designed to allow the separation of hydrogen and oxygen at different electrodes,with typical hydrogen evolution at the cathode and oxygen evolution at the anode ^[6]. But it is still a challenging task to develop a photoactive cathode for hydrogen production using materials based on earth-abundant elements ^[7].

As abundant,first row transition metal based complexes, cobaloxime derivatives have proved to be powerful catalysts for proton reduction into hydrogen with low overpotentials and have been successfully incorporated into several photochemical hydrogen generation systems using transition metal (Ru,Re,Ir,Pt,etc.) polypyridyl complexes,metal phorphyrins,organic dyes,or inorganic semiconductors as photosensitizers ^{[8, 9, 10]}. Although high turnover numbers based on molecular cobalt catalysts were achieved,sacrificial electron donors (triethylamine,ascorbate,etc.) were employed in the reported photo-water-reduction systems, which is impractical for real future applications. A technologically significant,sunlight driven hydrogen production system requires not only high efficiency but also minimal reliance on sacrificial donors. Therefore,attention has been paid to incorporation of cobaloxime derivatives into a photoactive cathode that could not only reduce the cobalt catalyst for water reduction under photoexcitation but also transfer the photogenerated holes to the corresponding anode for water oxidation ^{[11, 12]},for which hole transport in the photo active cathode is important. On the basis of p-type nano-structured NiO as the hole transport material in the photocathode of dye-sensitized solar cells,modification of dye sensitized p-NiO electrode by drop casting of molecular cobalt catalyst Co(dmgBF₂)₂(H₂O)₂ (dmgBF₂= difluoroboryldimethylglyoximate) has succeeded in splitting water under visible light ^[13]. In recent studies,quantum dot-sensitized NiO (CdS/NiO, CdSe/NiOetc.) cathodes exhibited faster hole transport than the dye sensitized NiO cathodes ^{[14, 15]},which motivates us to investigate the catalytic activity of the cobalt catalyst on the surface of quantum dot-sensitized NiO photocathodes. Herein,we present the integration of a molecular cobalt catalyst (1) [Co(dmgH)₂(4-Me-py)Cl] (dmgH = dimethylglyoximate,4-Me-py = 4-methylpyridyl) to CdS sensitized NiO film and its photoelectrocatalytic property in a PEC device for water splitting (Scheme 1).

Scheme 1

	Download: JPG larger image
Scheme 1.Closed photoelectrochemical cell,consisting of photocathode based on NiO film on FTO glass sensitized by CdS through TGA linker and then coated with cobalt catalyst1,Pt counter electrode and aqueous solution of Na2SO4(0.1 mol/L) as electrolyte.

Complex1was prepared according to an analogous synthesis in the literature ^[16],and its reduction potential was determined by the cyclic voltammograms recorded with a CHI-660E electrochemical potentiostat at a scan rate of 100 mV/s. The CV experiments were performed in a three-electrode cell under argon. The working electrode was a glassy carbon disk (diameter 3 mm),and the auxiliary electrode was a platinum wire. The reference electrode was an Agj Ag⁺ electrode (0.01 mol/L AgNO₃in CH₃CN). A solution of 0.05 mol/Ln-Bu₄NPF₆in CH₃CN was used as the electrolyte. The electrolyte solution was degassed by bubbling with dry argon for 10 min before measurement. The reduction potential is referenced to NHE (normal hydrogen electrode) according to the literature with Ferrocenium/Ferrocene couple as standard (0.40 Vvs.SCE and then SCEvs.NHE = 0.24 V) ^[17].

Nano-structured NiO powder was prepared according to the published procedure ^[18]. 0.3 g of the prepared NiO powder,1.05 g of terpineol,and 0.045 g of ethyl cellulose were suspended in 1.2 mL of ethanol. The mixture was stirred at room temperature for 2 days to prepare the screen printing paste. The paste was printed on FTO (fluorine-doped tin oxide) glass with an active area of 1cm² ,and the electrode was sintered at 450 ℃ for 2 h and then 550℃ for 30 min,resulting in NiO film with thickness of about 2mm. The NiO electrode was immersed in a solution of thioglycolic acid (TGA) in ethanol (0.02 mol/L) for 1 min and then dipped into a solution of Cd(NO₃)₂in methanol (0.1 mol/L) for 1 min. The electrode was rinsed with ethanol to get rid of excess Cd²⁺ ,after which the electrode was dipped into a solution of Na2S (0.1 mol/L) in a mixture of methanol/DI water (1:1,v/v) for 1 min. The electrode was finally rinsed with methanol and then dried in air. The process was repeated up to 5 times to get the desired NiO- TGA-CdS electrode.

The photocurrent measurements were carried out in a standard three-electrode cell with a CHI-660E potentiostat,using the above prepared photo-cathode as the working electrode,AgjAgCl AgCl (3 mol/L KCl aq.) as the reference electrode,and a platinum wire as the counter electrode. The cell was charged with an aqueous solution of Na₂SO₄(50 mmol/L) and flushed with argon for 15 minutes before photoreaction. A—0.2 Vvs.AgjAgCl potential was applied on the working electrode. The bias potential had been applied for two minutes before the illumination was carried out in order to stabilize the baseline and minimize the impact of background. The photocathode was illuminated using a light emitting diode array to give a light intensity of 250 W/m² calibrated by a OPHIR Nova II Laser power meter connected with a thermopile sensor (OPHIR,3A-P-FS). 3. Results and discussion

The transient photocurrent generated upon the on-off cycles of illumination on the CdS-NiO photocathode in the PEC is shown in Fig. 1. An initial current intensity of 17.5mA/cm² was obtained upon the first on-off cycle of illumination,indicating the catalytic water reduction by the electrons in the conduction band (CB) generated from light excitation of CdS and water oxidation by the holes in the valence band (VB) moved to the NiO film and the photoanode,respectively. When 5mL of the solution of 1 in acetonitrile (4 mmol/L) was deposited on the surface of CdS,the transient photocurrent generated from the first on-off cycle of the illumination increased to 25.0mA/cm² ,which was 1.43 times that obtained from the CdS-NiO photocathode. The photocurrent enhancement is attributed to the reduced recombination of the electrons and holes in CdS caused by the electron transfer from the conduction band of CdS to the cobalt catalyst 1. The above results are in accordance with the photocatalytic hydrogen production system composed of CdS and cobaloxime catalysts in the presence of a sacrificial electron donor in the mixture of H₂O/DMF,but NiO in the present PEC system could move the holes in the valence band of CdS to the anode for water oxidation,avoiding the use of sacrificial an electron donor. The stability of the hybrid photocathode was investigated after three on-off cycles of illumination and the photocurrent decay could be observed under long time irradiation. The increment of photocurrent intensity of the photocathode resulting from cobalt catalyst deposition could be observed clearly in the beginning of photoreaction but was not obvious after ten minutes (Fig. 2). Since similar cobalt catalysts have proven to function as catalysts for hydrogen production for several hours,the decomposition of cobalt catalyst 1 could be probably excluded. In order to get an insight into the reason for the photocurrent decay,the morphology and element composition of the photocathode surface before as well as after the photoreaction were revealed by scanning electron microscopy (SEM) and energydispersive X-ray analysis (EDX),respectively. The SEM images showed no obvious change on the surface of the photocathode before and after the photoreaction (Fig. S1 in Supporting information),suggesting that the structure of the CdS-NiO film remained the same. However,the cobalt element could be easily observed in the EDX spectrum before photoreaction but hardly found after photoreaction (Fig. S2 in Supporting information), indicating the loose contact of the cobalt catalyst with the photocathode caused by escaping from the electrode surface to water solution,leading to the decay of the photocurrent.

Fig. 1

	Download: JPG larger image
Fig. 1. The transient current responses to on-off cycles of illumination on photocathodes under an applied bias potential of—0.2 Vvs.AgjAgCl in the PEC.

Fig. 2

	Download: JPG larger image
Fig. 2. The photocurrent decays for photocathodes detected after three on-off cycles of illumination under an applied bias potential of—0.2 Vvs.AgjAgCl in the PEC.

The working mechanism of the present PEC device is proposed on the basis of the thermodynamic analysis summarized in Scheme 2,in which all the potentials were referenced to NHE for easy comparison. The cobalt catalyst 1 displays two reduction peaks in its cyclic voltammogram,with the first one at—0.32 V,assigned to the one electron reduction process of Co ^III /Co ^II ,and the other reduction event for Co ^II /Co ^I process occurring at—0.79 V. Based on the conduction band potential of CdS (—0.9 V) and inspired by the multiple electron transfer from the CdSe/ZnS core/shell quantum dot to a similar cobalt catalyst reported in the literature ^{[9, 10]},two electron reduction of 1 by the electrons from the conduction band of CdS is thermodynamically feasible. Besides,the valence band potential of NiO is located at 0.7 V ^[19],rendering hole injection from the valence band of CdS (1.5 V) also viable ^[14]. Therefore, upon visible light irradiation,electron-hole separation is generated in CdS,followed by multiple electron transfer from the conduction band of CdS to 1 to form Co ^I species. Protonation of the Co ^I species results in Co ^III -H hydride species that could go through the two molecular pathway to produce molecular hydrogen ^[9]. While the electrons accumulate in the conduction band of CdS,holes also accumulate in its valence band and move to the nano-structured NiO,which subsequently transfers to the counter electrode for water oxidation ^[13].

Scheme 2

	Download: JPG larger image
Scheme 2.The proposed mechanism based on redox potentials of the components of the photocathode.

A photoactive cathode initiated PEC water splitting system was developed based on earth abundant semiconductive material and cobalt catalyst. Under visible light irradiation,the photocurrent intensity of the PEC device increased by more than 40% after integration of 1 to the CdS-NiO photocathode. Although promising photocurrent increment was demonstrated by simple drop casting of cobalt catalyst to the CdS-NiO photocathode,photocurrent decay was observed and probably caused by the loose contact of the catalyst with the electrode surface. Therefore,studies on immobilization of cobalt catalyst on the surface of photocathode are in progress.

This work is supported by the Fundamental Research Funds for the Central Universities (No. HIT. IBRSEM. A. 201409),the Program for Innovation Research of Science in Harbin Institute of Technology (PIRS of HIT No. A201418,A201416),the National Natural Science Foundation of China (Nos. 21171044 and 21371040),and the National key Basic Research Program of China (973 Program,No. 2013CB632900).

Supplementary data associated with this article can be found, in the online version,at http://dx.doi.org/10.1016/j.cclet. 2014.09.011.