楊曉梅，吳 強(qiáng)，郭 茹，葉凱波，薛 屏，王曉中，賴小勇

（寧夏大學(xué)化學(xué)化工學(xué)院,省部共建煤炭高效利用與綠色化工國(guó)家重點(diǎn)實(shí)驗(yàn)室,銀川750021）

1 Introduction

Since the first report of photoelectrochemical water splitting on a TiO2electrode by Fujishima and Honda in 1972［1］，semiconductor-based photocatalysis driven by solar energy has become an intriguing approach for the economical and eco-friendly hydrogen production［2—4］，and numerous efforts have been devoted to the development of various highly active photocatalyst systems for water splitting［5—9］. TiO2and other wide bandgap semiconductors（such as ZnO，WO3，Ga2O3）could only utilize the parts（less than 5%）of solar beam in the UV region，which limited their practical application for solar hydrogen production. As a result，visible and near-infrared-light responsive semiconductor materials with relatively narrow band gaps have attracted more and more attention［10—18］. Cadmium sulfide（CdS）has a proper band gap for adsorbing visible light at wavelengths shorter than 516 nm and appropriate conduction band position for reducing water to generate hydrogen［19］，which is viewed as one of the most promising candidates for photocatalytic hydrogen production.However，separate common bulk CdS materials exhibited low photocatalytic activity and stability owing to the rapid recombination of photo-generated charges in bulk or on surface and serious photocorrosion during photocatalysis［20，21］. In principle，CdS materials on the nanoscale are expected to provide some unique advantages over their bulk counterparts，such as the increase of surface area and active sites，the reduction of the photogenerated-charge migrating distance from bulk to surface for photocatalytic reactions，the decrease of recombination probability. For example，Yuet al.［22］reported that CdS nanocrystals with a size ofca. 5 nm and high specific surface area ofca. 75.23 m2/g exhibited an obviously higher H2production performance than that for its aggregate with a large size of 30—100 nm and low specific surface area（ca.23.62 m2/g）. Wanget al.［23］also reported that the CdS nanorods with an average diameter ofca.10 nm exhibited a higher catalytic activity（200 μmol?h-1?g-1）for H2production than that（50 μmol?h-1?g-1）for the CdS nanospheres with a larger size of a few hundreds of nanometers. Zhanget al.［24］and Xianget al.［25］independently demonstrated that the ultrathin CdS nanosheet with a thickness ofca.4 nm exhibited a higher H2production rate than those for CdS aggregates and nanorods respectively. Moreover，loading proper cocatalysts（such as Pt，Au，Ag，Pd，Rh）on CdS materials has also been proved to be an effective strategy to improve the photocatalytic performance［26—30］.Domen and co-workers［31］synthesized an ordered array of CdS nanorods with diameters in 7—8 nm and a specific surface area of 73.6 m2/g and further loaded 10%（mass fraction）Pt as cocatalysts，which exhibited a 13 times improvement in photocatalytic H2production rate with respect to nude CdS nanoparticles. Considering the scarcity and high cost of noble metal-based cocatalysts，much attention has also been paid to exploring earth-abundant alternative，such as Ni- and Co-based cocatalysts［32—38］. Simonet al.［39］demonstrated that thein situphotodeposition of Ni nanoparticles（2—8 nm）on CdS nanorods could realize a significant enhancement of H2production rate under optional conditions. Xu and co-workers［40］reported that hydrothermally loading 1.2%（molar fraction）ultrafine NiS nanoparticles on CdS nanoparticles could reach a higher H2production rate than that for the 1%（mass fraction）Pt-loaded CdS photocatalysts. Recently，Jianget al.［41］reported that the catalyst prepared byin-situchemical deposition of amorphous NiS on CdS-based composite nanorods exhibited a 38 times improvement in visible-light-driven H2production rate with respect to pure CdS nanorods.Nevertheless，it is still necessary to design and develop advanced CdS-based photocatalysts with enhanced H2evolution performance.

In this work，we reported an ordered mesoporous cadmium sulfide（CdS）photocatalyst with both ultrathin crystalline frameworks ofca.5 nm and a large specific surface area of 238 m2/g，which could efficiently reduce the migration distance of photo-generated charge from bulk to surface in photocatalysis and provide more active sites for photocatalytic reaction. Furthermore，nickel sulfide as a cocatalyst was loaded on the surface of ordered mesoporous CdS by thein-situchemical deposition. Their photocatalytic performances for H2production in water were evaluated under visible light irradiation（λ≥420 nm）and a high photocatalytic H2evolution rate of 3.84 mmol?h-1?g-1could be reached after loading an optimal amount of NiS，which is 17.5 times that（0.22 mmol?h-1?g-1）for commercial CdS after loading the same amount of NiS.

2 Experimental

2.1 Synthesis of Ordered Mesoporous CdS

Ordered mesoporous CdS was synthesized through the nanocasting method using ordered mesoporous silica KIT-6 as hard template and cadmium thioglycolate as precursor，where KIT-6 was synthesized and aged at 40 °C according to the previous literature［42］and the cadmium thioglycolate was synthesized following the procedure previously proposed［43］. Typically，6.12 g of Cd（NO3）2·4H2O and 3.12 g of mercaptoethanol were dispersed in 5 g of water and kept at room temperature for 24 h. The resultant cadmium thioglycolate precursor｛Cd10［（SCH2CH2OH）］16（NO3）4｝was collected，washed with ethanol three times，and dried at room temperature overnight. Then，2 g of KIT-6 and 1.6202 g of Cd10［（SCH2CH2OH）］16（NO3）4were added to 20 mL of ethanol under stirring at 40°C. After the ethanol was completely evaporated，the resulting powder was heated up to 120°C at a heating rate of 1°C/min，and held at this temperature for 10 h. Thereafter，the temperature was increased to 160°C and kept for another 24 h to decompose Cd10［（SCH2CH2OH）］16（NO3）4into CdS. The resultant silica/CdS composite was treated several times at room temperature with concentrated alkaline solutions（2 mol/L NaOH）. Finally，the silica-free mesoporous CdS material was collected from the solution by centrifugation，washed with water and ethanol and dried at 70 °C for 12 h. For comparison，commercial CdS nanoparticles were purchased from Macklin Co.，Ltd.（purity of 98%，powder，Shanghai）.

Nickel sulfide as a cocatalyst was successfully loaded on the surface of ordered mesoporous CdS by thein-situchemical deposition［41］. Typically，50 mg of ordered mesoporous CdS was dispersed into a mixed aqueous solution containing 0.25 mol/L Na2SO3and 0.35 mol/L Na2S（80 mL）and then different amounts（3，5，20 or 500 μmol）of NiCl2was added and stirred for 10 min. The reaction system was kept at 6 °C，which could be directly used for photocatalytic H2production evaluation or further centrifuged，washed with water and dried at 70°C to get mesoporous NiS-loaded CdS material.

2.2 Characterization

X-Ray diffraction（XRD）pattern was recorded on a Bruker AXS D8 diffractometer（CuKα），operating under 40 mA and 40 kV. Transmission electron microscopy（TEM）and high-resolution transmission electron microscopy（HRTEM）images were recorded on an FEI Tecnai G2 F20 microscope operated at 200 kV.Samples for TEM observation were prepared by dispersing a small amount of powder in ethanol and then placing one or two drops of the suspension dropwise onto a holey carbon supported grid. N2physisorption was performed on a Micromeritics ASAP 2020 HD system at -196 °C and the pre-degassing condition was set at 150 °C for 13 h. Total pore volume was evaluated from the adsorbed volume at a relative pressure of 0.99.The specific surface area was calculated through the Brunauer-Emmet-Teller（BET）algorithm using the adsorption branch. The pore size distributions of the CdS samples were obtained through the Barrett-Joyner-Halenda（BJH）method using the desorption branches，whereas those of the KIT-6 sample were calculated by a nonlocal density functional theory algorithm.

2.3 Photocatalytic H2 Production Evaluation

The photocatalytic reactions for H2evolution were conducted on a commercial reaction system（CEL-PAEM-D8，Beijing China Education Au-light Co.，Ltd.，China）under visible-light irradiation［44］. A 300 W Xe-lamp（CEL-HXUV300）accompanied with an optical filter of 420 nm was used as the visible-light source to vertically irradiate the reaction vessel to generate hydrogen. The produced H2was detected every hour by an online gas chromatograph（GC-7920，N2as the carrier gas，Beijing China Education Au-light Co.，Ltd.，China）using a thermal conductivity detector（TCD）.

3 Results and Discussion

Fig.1 shows the low-angle XRD patterns of ordered mesoporous silica KIT-6 and ordered mesoporous CdS before and after removal of KIT-6. Three well-resolved diffraction peaks，indexed as the（211），（220）and（332）reflection of the 3D cubicIa3dsymmetry［42］，were observed in the XRD pattern of KIT-6（Fig.1 patterna）. After the formation of CdS within the channels of KIT-6，the intensity of diffraction peaks slightly decreased because of the weakened contrast of cubic mesostructures in the CdS/KIT-6 composite and an additional diffraction peak，indexed as the（110）reflection of the space groupI4132，appeared in the lower angle region，suggesting the occupation of CdS gyroid frameworks within only one of two sets of double gyroid mesopores in KIT-6 and possibly forming a single uncoupled subframework［45，46］. After removal of KIT-6，the（110）diffraction peak remained in the XRD pattern of the resultant ordered mesoporous CdS，further confirming the above-mentioned deduction.

Fig.1 Low-angle XRD patterns of ordered mesoporous silica KIT-6(a),CdS/KIT-6(b)and ordered mesoporous CdS after removal of KIT-6(c)

Fig.2 Wide-angle XRD patterns of ordered mesoporous CdS after removal of KIT-6

The wide-angle XRD pattern of the resultant ordered mesoporous CdS is shown in Fig.2. The three intense diffraction peaks in the range of 10°—85°could be indexed as the（111），（220），and（311）Bragg reflections of CdS phase（JCPDS No. 89-0440）［47］. The broadening of highly intense diffraction peaks reveals the nanoscale structure and crystallinity of the CdS subframework. There is no other diffraction peaks ascribable to impurities such as CdO，which usually appears at higher crystallization temperature owing to the partial oxidation of CdS by the ambient oxygen［31］.

N2adsorption-desorption isotherms reveal that the ordered mesoporous silica KIT-6 had a specific surface area of 810 m2/g（Fig.3 isotherma），whereas the resultant ordered mesoporous CdS exhibited a specific surface area of 238 m2/g（Fig.3 isothermb），significantly higher than those for common commercial CdS（81 m2/g，F(xiàn)ig.3 isothermc）and ordered mesoporous CdS previously reported in literatures（73.3—120 m2/g）［31，48—50］. It should be reasonable if considering the thickness of CdS subframework confined within the small mesochannels of 5.6 nm for silica template KIT-6（Fig.4 curvea）could be effectively limited toca. 5 nm by the pore walls of KIT-6，which would also be confirmed by its TEM image［Fig.5（A）］. The corresponding pore size distribution curve of mesoporous CdS（Fig.4 curveb）exhibited two types of mesopores centered at 5.1 and 10 nm respectively. The former is consistent with the pore wall thickness of KIT-6，whereas the latter corresponds to the uncomplete occupation of CdS within two sets of mesopores in KIT-6，which is in agreement with its low-angle XRD pattern. Both high specific surface area and large mesopores of the resultant ordered mesoporous CdS would be especially benefical for providing more active sites and enhancing the mass-transfer，resulting in a significantly improved catalytic performance.

Fig.3 Nitrogen adsorption-desorption isotherms of ordered mesoporous silica KIT-6(a),ordered mesoporous CdS after removal of KIT-6(b)and commercial CdS(c)

Fig.4 Pore size distribution curves of ordered mesoporous silica KIT-6(a) and ordered mesoporous CdS after removal of KIT-6(b)

Fig.5 TEM image of ordered mesoporous CdS without NiS(A),HRTEM images of ordered mesoporous CdS without(B) and with(C) NiS, HADF STEM image of ordered mesoporous CdS with NiS(D1)and the corresponding elemental maps for Cd,Ni and S(D2—D4)

Typical TEM image for mesoporous CdS products is shown in Fig.5（A）. From the TEM image，it could be clearly observed that these mesoporous CdS are mainly composed of some particles with periodically mesostructure. The corresponding high resolution TEM（HRTEM）image［Fig.5（B）］of the resultant ordered mesoporous CdS shows the ultrathin subframeworks with the thickness of 4—5 nm，which is comparable to the pore size of KIT-6. Uncoupled subframeworks with large pores of 10 nm can be observed in the edge of particles，which confirms that the resultant ordered mesoporous CdS replicated only one of two sets of enantiomeric mesochannels in KIT-6 in some local domains，in line with the low-angle XRD data（Fig.1）. The clear lattice fringes with a lattice spacing of 0.336 nm，which corresponded to the（111）lattice plane of CdS，also confirm the high crystallinity of the CdS subframeworks. These results suggest a relatively successful replication from KIT-6 to ordered mesoporous CdS with ultrathin and crystalline frameworks. Afterin-situchemical deposition of NiS，the mesoporosity and crystalline ultrathin frameworks of ordered mesoporous CdS remain well［Fig.5（C）］. The corresponding high angular dark field（HADF）scanning TEM（STEM）image［Fig.5（D1）］and elemental maps for Cd，Ni and S［Fig.5（D2—D4）］confirms the uniform distribution of NiS on the surface of ordered mesoporous CdS and no significant agglomeration is observed.

Fig.6 Time courses of H2 evolvation rates for ordered mesoporous CdS(50 mg) loaded with different amounts of NiS

Fig.7 Average H2 evolvation rates for ordered mesoporous CdS(50 mg)loaded with different amounts of NiS

Photocatalytic H2production evaluations were conducted in an 80 mL aqueous solution containing 50 mg of the resultant ordered mesoporous CdS photocatalysts loaded with 3—500 μmol of NiS cocatalyst byin-situchemical deposition with sacrificial reagents of Na2SO3and Na2S under the visible light（λ≥420 nm）irradiation of a 300 W Xe lamp. For comparison，commercial CdS nanoparticles with a specific surface area of 81 m2/g were also evaluated，keeping all other testing conditions the same. As shown in Fig.6 curvea，the ordered mesoporous CdS loaded with 3 μmol of NiS exhibits a H2evolution performance of 15.9 mmol/g under visible light irradiation for 5 h. Increasing the loading amount of NiS to 5，20 and 500 μmol，the H2evolution performance under visible light irradiation for 5 h correspondingly increase to 19.2 mmol/g，and then decrease to 9.4 and 0 mmol/g（Fig.6 curvesb—d）. Fig.7 shows that the average H2evolvation rates for ordered mesoporous CdS loaded with 3—500 μmol of NiS and commercial CdS loaded with 5 μmol of NiS are 3.18，3.84，0.85，0 and 0.22 mmol?h-1?g-1，respectively. These results indicate that an optimal amount of NiS can highly promote the H2production performance in ordered mesoporous CdS，which could be explained by the mechanism proposed by Xu and co-workers［40］that NiS cocatalyst could not only efficiently improve the separation and migration of photogenerated electrons from the conduction band（CB）of CdS to NiS，but also promote the adsorption and desorption kinetics for H2evolution. However，excessive NiS would possibly shield the light absorption of mesoporous CdS and cover the active sites［41］，resulting in the decrease of photocatalytic H2production performance and even the complete loss of the catalytic activity. Notably，commercial CdS nanoparticles loaded with the same amount（5 μmol）of NiS exhibits an extremely low H2evolution performance of 1.1 mmol/g under visible light irradiation for 5 h（Fig.6 curvee），which is 17.5 times less than that for the resultant ordered mesoporous CdS loaded with 5 μmol of NiS，although the specific surface area of the former is only 2.9 times less than that of the resultant ordered mesoporous CdS. That may be due to that the ultrathin subframework ofca.5 nm in the resultant ordered mesoporous CdS could significantly shorten the migration distance of photo-generated charges from bulk to surface in photocatalysis and thus reduce the recombination of the photo-generated electrons and holes，thus resulting in more effective photo-generated electrons for H2evolution. After three reaction runs（Fig.S1，see the Supporting Information of this paper），the sample retained 96.4% of the initial photocatalytic activity under visible-light irradiation and reached a stable H2evolution rate of 3.70 mmol?h-1?g-1，superior over 10%（mass fraction）Pt-loaded CdS nanowire arrays（ca. 2.6 mmol?h-1?g-1）［31］. Moreover，a substantial difference between mesoporous CdS and commercial nonporous CdS should be noted，where thein-situchemical deposition of NiS may partially lead to the pore block of mesoporous CdS and decrease the specific surface area（ca. 118 m2/g，F(xiàn)ig.S2，see the Supporting Information of this paper），suggesting that there should be still a great potential to further enhance photocatalytic H2production performance of the ordered mesoporous CdS if improving NiS deposition without the loss of specific surface area.

4 Conclusions

We have synthesized an ordered mesoporous CdS photocatalyst through the nanocasting method using ordered mesoporous silica as hard template，which is constructed with ultrathin crystalline frameworks ofca.5 nm and possesses a large specific surface area of 238 m2/g. The resultant ordered mesoporous CdS with both high specific surface area and ultrathin crystalline frameworks exhibited an excellent photocatalytic H2evolution activity after loaded with an optimal amount of NiS，significantly superior to commercial CdS because of more active sites derived from larger surface area and more effective photo-generated electrons for H2evolution resulted from the shorten migration distance of photo-generated charges from bulk to surface.More investigations on the bulk composition adjustment and surface functionalization of such an ordered mesoporous CdS with both high specific surface area and ultrathin crystalline frameworks may be worthy for further improving its photocatalytic H2evolution.

The Supporting Information of this paper see http：//www.cjcu.jlu.edu.cn/CN/10.7503/cjcu20210035.

This paper is supported by the National Natural Science Foundation of China（No. 52062043）.