One-step chemical bath co-precipitation method to prepare high hydrogen-producing active ZnxCd1-xS solid solution with adjustable band structure

A series of ZnxCd1−xS solid solutions were synthesized through a simple one-step chemical bath co-precipitation route. The microstructure, morphology, composition and optical properties had been thoroughly investigated. The results showed that the as-obtained ZnxCd1−xS samples exhibited monodisperse spherical feature. These monodisperse ZnxCd1−xS spheres were composed of a large number of 5–10 nm crystal grains. The chemical composition of the ZnxCd1−xS solid solutions can be controlled by adjusting the ratio of Zn source to Cd source and the dosage of ammonia solution. Meanwhile, the energy band structure and photocatalytic properties can be optimized. The photocatalysis experiment results revealed that the as-synthesized Zn0.30Cd0.70S sample exhibited optimum water splitting performance. The satisfactory hydrogen evolution rate (HER) of Zn0.30Cd0.70S reached 27.004 mmol g−1 h−1 even without any cocatalysts, which was more than 48 times that of pure CdS (0.561 mmol g−1 h−1). This enhancing effect can owe to the balance between light absorption capacity and redox potential caused by the incorporation of Zn in the ZnxCd1−xS solid solutions.


Introduction
As human society is constantly evolving, energy crisis and environmental damage problems have become the two major issues that must be solved immediately [1,2]. Hydrogen energy, a clean and renewable fuel, has received extensive attention. Since Honda and Fujishima discovered a feasible photoelectrochemical (PEC) method for reducing water to produce hydrogen on the TiO 2 electrodes [3], this photocatalytic technology of using solar energy to split water has received widespread attention by researchers. Because this PEC method is a promising green solution for converting solar energy into chemical energy to produce hydrogen, which is considered to be the most ideal approach to solving both environmental and energy problems [4][5][6][7]. Since then, numerous potential semiconductor materials have been used in the practice of PEC water splitting, such as SrTiO 3 [8][9][10], BaTiO 3 [10,11], ZnO [12,13], CeO 2 [14,15], g-C 3 N 4 [15][16][17][18], ZnS [19][20][21] and CdS [21][22][23][24]. However, most of them can only operate in the ultraviolet region, which greatly limits their wide applications. Hence, it is imperative to develop visible light-responsive catalysts that can more effectively use solar energy.
As a typical semiconductor photocatalyst that can respond to visible light, CdS has a small band gap (E g ) and a favorable redox potential, which is suitable for photocatalytic water splitting [25]. Unfortunately, the severe recombination process of photogenerated carriers and serious photocorrosion result in low photocatalytic performance and poor cycle stability [23][24][25], which severely limit the wide application in photocatalytic water splitting. At the same time, CdS has strong toxicity. On the contrary, as another representative of II-VI semiconductor, ZnS has excellent properties complementary to CdS, such as high energy conversion efficiency, good photocatalytic stability, low toxicity and relatively negative reduction potential, although it only works under ultraviolet light. Therefore, the combination of ZnS and CdS should be a viable method to take advantage of the virtues of these two semiconductors and improve their respective defects. Constituted ternary metal sulfides system (Zn x Cd 1-x S), which can continuously regulate the E g from about 2.4 eV (CdS) to 3.8 eV (ZnS) by adjusting the content of Zn and Cd ions [26][27][28], often shows excellent catalytic performance. Up till the present moment, many efforts have been paid to synthesize high-performance Zn x Cd 1-x S catalysts. For example, Chen et al. had prepared H-ZnCdS rhombic cages, fabricated by vulcanizing the single precursor of ZIF-8 and then performing cation exchange by hydrothermal method at 160°C [29]. Li et al. had developed a route to directly pyrolyze zinc-cadmium-thiourea complex at 500°C to obtain high-efficiency Zn 1-x Cd x S photocatalysts [27]. Zhang et al. used layered porous ZnS nanospheres as a sacrificial template, followed by co-hydrothermal with Cd(Ac) 2 at 140°C to form Zn x-Cd 1-x S double-shell hollow nanospheres [30]. In addition, Xiang et al. used oleylamine as an organic solvent to successfully synthesize sphalerite-type structured Cd 1-x Zn x S nanocrystals through an organic solution method at 280°C [31]. From the results, these Zn x Cd 1-x S solid solutions obtained through the above strategy showed significantly improved H 2 evolution performance compared with pure CdS materials. But, the above-reported synthesis process usually involves complex routes and stringent reaction conditions, such as high temperature, high pressure, multiple precursors or organic solvents. Hence, it is meaningful to open up a simple and gentle route to synthesize Zn x-Cd 1-x S photocatalytic materials and at the same time aim to further improve their hydrogen evolution performance. Among many material synthesis methods, chemical bath method is widely used in the synthesis of various materials because of its simple and convenient operation [32]. The chemical bath method can be carried out at low temperature and normal pressure, and the reaction time is generally short. Therefore, for the consideration of scale production and practical application, this method has obvious advantages. For this reason, we tried to synthesize the Zn x Cd 1-x S photocatalyst under a mild condition through a simple one-step chemical bath co-precipitation method. To our knowledge, there is no report on the synthesis of monodisperse Zn x Cd 1-x S nanosphere material for photocatalytic hydrogen evolution by this method.
Here, we successfully synthesized a series of Zn x-Cd 1-x S solid solution materials via a simple chemical bath co-precipitation route without using any organic solvents or templates. These solid solutions obtained by inserting Zn ions into the CdS lattice inherit the advantages of the visible light response of CdS and the high reduction potential of ZnS and improve the problems of severe photocorrosion of CdS and the narrow light response range of ZnS. The resulting photocatalysts were characterized by various methods and applied to water reduction. In addition, we also explored the effect of ammonia solution on the formation of solid solutions.

Experimental section Materials
Zinc nitrate hexahydrate, cadmium nitrate tetrahydrate and thiourea were obtained from Aladdin. Trisodium citrate dihydrate and ammonia solution (NH 3 ÁH 2 O, 25%) were obtained from Sinopharm. All raw materials were of analytical grade. Deionized water from local sources was used as solvent.

Synthesis of Zn x Cd 1-x S and Zn x Cd 1-x S-n monodisperse spheres
Samples were synthesized using a simple one-step chemical bath co-precipitation method. Typically, to synthesize Zn 0.30 Cd 0.70 S (Zn 0.30 Cd 0.70 S-1) solid solution material, 5 mL of trisodium citrate dihydrate (0.12 M), 7 mL of cadmium nitrate tetrahydrate (0.05 M) and 3 ml of zinc nitrate hexahydrate (0.05 M), 1 mL of ammonium solution (17.03 M), and 4 mL of thiourea (0.5 M) were sequentially poured into a 250-ml three-necked flask containing 100 mL of deionized water every 10 min under continuous magnetic stirring. Then, the reaction system was kept in an oil bath at 80°C for 3 h. Subsequently, the suspended yellow precipitate was collected and preserved. By appropriately regulating the ratio of Cd source to Zn source (the amount of ammonia solution was 1 ml) or the dosage of ammonia solution (the ratio of Cd source to Zn source was 7 to 3), a series of Zn x Cd 1-x S and Zn x Cd 1-x S-n (the amount of ammonia solution was n ml) solid solution materials can be easily obtained via a similar procedure.

Characterization
Crystal structure and crystalline quality of all specimens were examined by the powder X-ray diffraction (D/max 2550, Rigaku Corporation) in the 2h = 108-808 range. The morphologies and structures were characterized by scanning electron microscope (Nova Nano SEM 230, FEI Electron Optics B.V) and transmission electron microscope (JEM-2100F, Japan Electronics Co. LTD). Microanalysis of energy-dispersive X-ray (EDX) spectrum was also obtained from the above-mentioned scanning electron microscope. Element mapping is performed on transmission electron microscope (Titan G260-300). The surface element information was collected from X-ray photoelectron spectroscopy (K-Alpha 1063, Thermo Fisher). The UV-Vis diffuse absorption spectra were gained from a UV-Vis spectrometer (Evolution 220, Thermo Fisher Scientific) to reveal the optical physical properties.

Photocatalytic tests
The catalytic ability of all synthesized solid solution catalysts was evaluated by reducing water under simulated sunlight irradiation. In a representative catalytic hydrogen production reaction, Zn x Cd 1-x S photocatalyst (10 mg) and sacrificial reagents (0.035 mol of Na 2 S and 0.025 mol of Na 2 SO 3 ) were ultrasonically dispersed in a glass reactor containing 0.1 L of deionized water. Then, the reactor was connected to a vacuum system so that the entire reaction process was performed in a vacuum state. The above reactor was irradiated under a 300 W Xe lamp and kept at 5°C by circulating condensed water. The amount of hydrogen produced by the reaction system was analyzed by a gas chromatograph every hour.

Electrochemical analysis
The CHI 660E electrochemical workstation was used to reveal the PEC properties of the prepared catalysts, including Mott-Schottky curves, electrochemical impedance spectroscopy (EIS), and transient-state photocurrent response. A standard three-electrode system was adopted. Saturated Ag/AgCl electrode was acted as reference electrode, while platinum foil was used as counter electrode. Working electrodes were obtained as follows: First, catalyst specimen (5 mg) was ultrasonically dispersed in the glycol solution (2 mL) and kept under magnetic stirring for 24 h; then, the catalyst slurry (100ul) was evenly coated on the fluorine-doped tin oxide (FTO) glass (10 mm 9 10 mm); finally, the working electrodes can be obtained by completely evaporating the solvent. The EIS test was performed in the frequency range of 1000 kHz-0.1 Hz with an amplitude of 5 mV. Under simulated sunlight, transient photocurrent was collected for a period of 40 s. The Mott-Schottky curves were recorded at a frequency of 500 Hz and an amplitude of 50 mV under the dark condition. The applied voltage varies from -1.5 V to 1.0 V.

Results and discussion
Powder X-ray diffraction (XRD) was used to analyze the crystal structure and crystalline quality of assynthesized Zn x Cd 1-x S products (Fig. 1a). As seen from Fig. 1a, black line, the XRD pattern of pure CdS sample can be assigned to the hexagonal wurtzite structure, which could be indexed as the standard card (JCPDS no. 41-1049). The three diffraction peaks located at 24.8°, 26.5°and 28.2°corresponded to (100), (002) and (101) planes, respectively. It was not difficult to find that as the value of x increased, the pattern peaks of Zn x Cd 1-x S shifted to a large angle. This phenomenon indicated that zinc ions had been incorporated into the crystal lattice of CdS crystals to form a solid solution, while reducing the lattice distance of CdS crystals [30,33,34]. What's more, the two diffraction peaks near 24.88 and 28.28 were not particularly noticeable due to poor crystallization performance induced by low temperature reaction and the increase in zinc composition. And the absence of additional phases meant that the synthetic material was a single phase.
The XRD patterns of Zn x Cd 1-x S-n (attained with the addition of n mL NH 3 ÁH 2 O, while the ratio of Cd source to Zn source was 7 to 3 in the initial reaction solution) are revealed in Fig. 1b. The XRD peaks not only shifted to left side with the increase in the n value, but also the two diffraction peaks near 24.88 and 28.28 gradually became prominent. The shift of the diffraction peaks to the left meant that the cadmium content had increased in the solid solution. At the same time, the observation of the two XRD peaks near 24.88 and 28.28 showed the improvement in the crystallinity of the obtained material. These phenomena strongly proved that the dosage of ammonia solution not only adjusted the composition of Zn x-Cd 1-x S, but also affected the crystallinity.
To reveal the optical properties of synthetic materials, solid UV-Vis diffuse reflectance spectra (DRS) are given in Fig. 2a. Obviously, as the zinc content increased, the absorption edges of all specimens showed successive blue shift one after another instead of two-stage absorption, which strongly confirmed the formation of solid solutions. The result was consistent with the XRD result. This also agreed with other studies [27,28,33,34]. To further reveal the changes in the optical properties, the E g energy values corresponding to Zn x Cd 1-x S (x = 0, 0.14, 0.30, and 0.46) materials were determined by Kubelka-Munk method [20,27,35]. As seen from Fig. 2b, the estimated E g values were 2.29, 2.39, 2.48 and 2.56 eV, respectively. It showed that the incorporated Zn ions in the CdS crystal increased the E g , thereby reducing the incident light absorption range of Zn x Cd 1-x S catalysts.
To achieve a further insight of the energy band structure about the as-synthesized Zn x Cd 1-x S materials, we performed Mott-Schottky measurements [36]. The Mot-Schottky curves (Fig. 2c) showed that the reduction potential of all solid solutions had a negative correlation with x value. A larger zinc content results in a more negative conduction band (CB) edge, and a more negative CB edge means stronger reducing ability, which is helpful to strengthen the activity of reduction to produce H 2 [28,30]. From the results of Mott-Schottky measurements and Kubelka-Munk plots, the band structure information of Zn x Cd 1-x S samples can be obtained in detail, as shown in Fig. 2d and Table 1.
The morphologies and structures of the synthetic products with different metal compositions were detected by scanning and transmission electron microscopy (SEM and TEM). As seen from Fig. 3a-b and Fig. 4a-f, all prepared Zn x Cd 1-x S samples were uniformly monodisperse nanospheres. The surface of Zn x Cd 1-x S nanospheres tended to be regular as Zn ions were continuously incorporated into the CdS crystal lattice. TEM micrographs are shown in Fig. 3b and Fig. 4d-f, indicating that the rough surface of the nanospheres is composed of numerous small-sized nanocrystals. As shown in Fig. 3c, the high-resolution TEM micrograph of Zn 0.30 Cd 0.70 S revealed that the grain size was about 5-10 nm, which is supposed to contribute to a significant increase in catalytic performance [29]. The d spacing of lattice fringes of the random nanoparticle was measured to be 3.36 Å and 2.07 Å corresponding to the (002) and (110) planes of the Zn 0.30 Cd 0.70 S solid solution, respectively, which matched the XRD results. The EDX spectrum (Fig. 3d) confirmed that the materials we prepared were composed of cadmium, zinc and sulfur elements. And the elemental compositions of all samples were determined, as shown in Table 1. As shown in Fig. 3e-h, the element mapping images were given, which confirmed that Zn, Cd and S elements were uniformly distributed throughout the as-prepared monodisperse nanospheres.
As shown in Fig. 4g-i, the morphologies and size distribution of Zn x Cd 1-x S-n specimens were displayed at different ammonia levels. It was very clear that the dosage of ammonia only affected the size of the sample, but the spherical feature did not change significantly. With the increase in ammonia solution, the sample size increased from approximately 80 to 148 nm. At the same time, the EDX test results (Table 1) showed changes in the composition of the prepared samples, which confirmed the conclusion of XRD. In the case where the Zn/Cd precursor solution ratio remained unchanged, proper adjustment of the dosage of ammonia solution could change the composition of the final products. With the increase in the dosage of ammonia solution, thiourea was accelerated to decompose to produce S 2-, which significantly affected the nucleation and growth process, leading to accelerated precipitation reaction. In this process, on the one hand, the grain growth was promoted and the grain size became larger, which led to an increase in crystallinity; on the other hand, the size of the solid solution spheres increased with the acceleration of the precipitation reaction. As we all  know, under the same conditions, the solubility constant of ZnS is significantly higher than that of CdS. Therefore, CdS is easier to deposit. So, during the co-precipitation process, Cd 2? ions were obviously in a favorable position during the reaction with S 2ions, which led to a slight increase in the Cd content in the resulting solid solution [37]. The surface element information of photocatalysts was detected by X-ray photoelectron spectroscopy (XPS). Figure 5a is XPS full spectra of the corresponding Zn x Cd 1-x S samples. Obviously, the XPS full spectrum of CdS was mainly composed of the diffraction peaks of cadmium and sulfur. In addition, XPS peaks of cadmium, zinc and sulfur can be discovered in the other three Zn x Cd 1-x S samples. Highresolution XPS spectra did prove the presence of cadmium, zinc and sulfur (Fig. 5b-d). The results were completely consistent with the EDX test. As the zinc content increased, the spectral intensity of Zn 2p gradually increased (Fig. 5c), while the spectral intensity of Cd decreased slightly (Fig. 5b), denoting that the relative content of zinc and cadmium can be easily adjusted [38]. The S 2p spectra of all specimens showed similar peak intensities (Fig. 5d). As the zinc content increased, the binding energy of S 2p gradually increased from 161.28 eV of CdS to 161.78 eV of Zn 0.46 Cd 0.54 S, indicating that a large number of more stable ZnS chemical bonds were formed in Zn x Cd 1-x S samples [38]. At the same time, the binding energy of Cd 3d 5/2 and Cd 3d 3/2 increased from 405.14 eV and 411.04 eV to 405.36 eV and 412.09 eV. Continuous changes in the binding energy also suggested the formation of solid solutions.
The catalytic performance of synthetic Zn x Cd 1-x S specimens was evaluated toward H 2 evolution without the assistance of any cocatalyst. Mixed aqueous solutions containing 0.25 M SO 3 2and 0.35 M S 2were used as sacrificial agents system to prevent photocorrosion of the sulfide photocatalyst. As seen from Fig. 6a, the hydrogen evolution rate (HER) of pure CdS under simulated sunlight irradiation was negligible (0.561 mmol g -1 h -1 ).
-Surprisingly, with the embedding of Zn ions, the HER of Zn x Cd 1-x S increased dramatically. As for Zn 0.3 Cd 0.7 S, the HER reached a peak value of 27.004 mmol g -1 h -1 , which was about 48 times the HER of pure CdS. However, with the further increase in zinc content, the performance of Zn 0.46 Cd 0.64 S catalyst declined. From the DRS and Mott-Schottky tests results, it can be known that as the zinc content increased, the photocatalyst sacrificed part of the visible light utilization rate, and at the same time, a more negative CB edge was obtained. The light-harvesting capacity and redox potential of a photocatalyst are crucial and significantly affect the photocatalytic hydrogen production activity [27,38]. Here, when x \ 0.3, the catalytic performance mainly depends on the redox ability of the photocatalyst; in contrast, when x [ 0.3, the performance is mainly determined by the photocatalytic light absorption capacity. Properly adjusting the relative content of zinc and cadmium in the Zn x Cd 1-x S solid solution, the light-harvesting ability and the redox potential have reached a good balance, which is the main reason why Zn 0.3 Cd 0.7 S shows the best catalytic performance. The catalytic performances of Zn x Cd 1-x S-n samples are also given in Fig. 6b. The catalytic performance was similar to that of Zn x Cd 1-x S. It was worth noting that the catalytic performance of Zn 0.11 Cd 0.89 S-4 dropped significantly, even lower than that of CdS. This is because the size of the obtained Zn 0.11 Cd 0.89 S-4 sample increases significantly with the increase in the ammonia solution dosage.
The stability of the Zn 0.3 Cd 0.7 S specimen was tested in the identical catalytic process. Figure 6c reveals the brilliant performance of the Zn 0.3 Cd 0.7 S sample during the four cycles. After 16 h of irradiation, Zn 0.3-Cd 0.7 S specimen still maintained satisfactory hydrogen production performance. In addition, the catalyst after the cycle test was recovered and further characterized by XRD (Fig. 7a) and TEM (Fig. 7b). It was affirmed that the basic crystal structure and monodisperse spherical characteristics of Zn 0.3 Cd 0.7 S were retained after four cycles. All the above phenomena have strongly proved the excellent photocatalytic hydrogen production performance of Zn 0.3 Cd 0.7 S.
To further reveal the PEC properties of the synthetic Zn x Cd 1-x S solid solutions, electrochemical impedance spectroscopy (EIS) and transient photocurrent response measurements were carried out. As depicted in Fig. 8a, EIS Nyquist plots, the arc radius of Zn 0.30 Cd 0.70 S was smallest among all samples. Therefore, the as-obtained Zn 0.30 Cd 0.70 S material had higher charge transfer efficiency and lower charge recombination possibility than other samples [30,39]. This is because an appropriate amount of doped zinc in the CdS lattice can provide a suitable impurity level, so the excited electrons from the CdS valence band can be easily injected into the conduction band [40]. The photocurrent response of Zn x Cd 1-x S specimens is given in Fig. 8b. These catalysts exhibited a very stable and strong photocurrent response to simulated solar radiation. When the incident light was shielded, the current density quickly decayed to near zero. Apparently, for Zn 0.30 Cd 0.70 S sample, the observed photocurrent density was significantly higher than that of other samples, which indicated that there were more separated photogenerated electrons and holes in the catalyst system. This can be attributed to the effective light utilization efficiency and fast photogenerated carrier mobility. It was worth noting that the decay trend of photocurrent and hydrogen production capability was synchronized. So, the improved photocatalytic hydrogen production performance should also be attributed to the improved photogenerated charge transfer capacity and elevated photocurrent.
Generally, an effective photocatalytic reaction involves three processes, including light harvesting, carrier generation and migration, and surface redox reaction. As shown in Fig. 9, the photocatalytic hydrogen production reaction occurring on Zn 0.30-Cd 0.70 S photocatalyst is revealed. Under the excitation of sunlight (the energy of this part of the incident photon is greater than or equal to the E g of the Zn 0.30 Cd 0.70 S), electrons are excited from VB and  migrated to CB, generating negatively charged highly active electrons on the CB of the Zn 0.30 Cd 0.70 S, while leaving positively charged holes in the VB. Because there are a large number of sacrificial agents (Na 2 S and Na 2 SO 3 ) with strong reducing ability in the catalytic reaction system, the photogenerated holes would be captured and consumed quickly. The photogenerated electrons could reach to the surface reaction site and undergo a reduction reaction with the adsorbed water to produce hydrogen. Although the E g of the Zn x Cd 1-x S materials are broadened due to the insertion of Zn ions in the CdS structure, the visible light-harvesting capacity decreases. But at the same time, the obtained solid solution materials photogenerated carriers gain stronger redox ability, which is beneficial to the rapid progress of the electron catalyzed hydrogen production reaction and the hole oxidation reaction. The rapid consumption of electron-hole pairs is also conducive to the regeneration of photogenerated carriers and the continuous and efficient progress of catalytic reactions. The rapid transport and transfer of carriers had also been confirmed by EIS and transient photocurrent tests. The best hydrogen production performance was obtained on Zn 0.30 Cd 0.70 S sample, indicating that the improvement in photocatalytic performance benefits from the balance between light absorption and redox potential, as well as rapid carrier transfer and improved photocurrent response.

Conclusions
In summary, by controlling the insertion of Zn into the CdS lattice, a series of monodisperse Zn x Cd 1-x S nanospheres were successfully synthesized via a simple chemical bath co-precipitation route. Through various characterizations and testings, the as-prepared Zn 0.30 Cd 0.70 S material exhibited the best water splitting activity with a maximum hydrogen evolution rate of 27.004 mmol g -1 h -1 , which was about 47 times higher than that of pure CdS (0.561 mmol g -1 h -1 ). Moreover, the Zn 0.30 Cd 0.70 S sample exhibited excellent catalytic stability under long-term light irradiation. This enhancing effect can owe to the balance between light absorption and redox potential caused by the incorporation of Zn in the solid solution. Interestingly, by adjusting the dosage of the ammonia solution, we found that the ammonia concentration during the reaction was positively correlated with the cadmium content, crystallinity and size of the obtained Zn x Cd 1-x S solid  solution samples. This work provides new possibilities for the simple preparation of Zn x Cd 1-x S photocatalysis with high activity and stability and emphasizes the significant effect of ammonia on the formation of Zn x Cd 1-x S solid solution.