The influence of the sacrificial agent nature on transformations of the Zn(OH)2/Cd0.3Zn0.7S photocatalyst during hydrogen production under visible light

Photocatalysts based on zinc hydroxide and a solid solution of CdS and ZnS were prepared via the precipitation method and used for photocatalytic hydrogen production from aqueous solutions of inorganic (Na2S/Na2SO3) and organic (ethanol) sacrificial agents. The photocatalysts were tested in cyclic experiments for hydrogen evolution and studied using X-ray diffraction (XRD), UV-Vis diffuse reflectance spectroscopy, high-resolution transmission electron microscopy (HRTEM), energy-dispersive X-ray spectroscopy (EDX), and X-ray photoelectron spectroscopy (XPS) techniques. Different transformations of the β-Zn(OH)2 co-catalyst were observed in the presence of inorganic and organic sacrificial agents; namely, ZnS was formed in Na2S/Na2SO3 solution, whereas the formation of ε-Zn(OH)2 was detected in solution with ethanol. The composite Zn(OH)2/Cd1−xZnxS photocatalysts have great potential in various photocatalysis processes (e.g., hydrogen production, CO2 reduction, and the oxidation of organic contaminants) under visible light.


Introduction
Photocatalytic hydrogen evolution is known to be an environmentally friendly approach to the synthesis of H 2 for various energy applications. 1,2 From a practical point of view, the photocatalysts for hydrogen production should be active under visible light, because the solar radiation spectrum has ca. 40% of photons in the visible region of the spectrum. 3 The materials based on CdS, 4 ZnS, 4 and their solid solution Cd 1Àx Zn x S (ref. 5) have great promise for this purpose. Unfortunately, the activity and stability of the Cd and Zn sulde photocatalysts are quite low owing to the intense recombination of charge carriers and photocorrosion. The deposition of co-catalysts, such as oxides, 6-9 hydroxides 10-12 or other suldes, [12][13][14][15][16][17] can substantially improve the target parameters of the photocatalyst and enhance its activity. The photocatalysts based on the composition of Cd 1Àx Zn x S and Zn(OH) 2 show a higher activity and stability compared to Cd 1Àx Zn x S alone and are attractive for the photocatalytic hydrogen evolution from water-alcohol solutions. 18 In the 1950s, G. K. Boreskov stated that the chemical and phase composition of the heterogeneous catalyst was changed during the reaction owing to its interaction with the reagents. 19 Studying the effect of the reagent nature on the catalyst composition and its catalytic activity is an important task from a fundamental point of view to discover the mechanism of the catalytic action. In a previous study, 18 we demonstrated that the activation of the Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalyst occurs during the photocatalytic hydrogen production from both the Na 2 S/Na 2 SO 3 and ethanol aqueous solutions. This study aims to investigate in detail the changes in the composition of the Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalyst during its activation in this process and to establish the role of the sacricial agent on the transformation of the photocatalyst. In this paper, we show that for the Na 2 S/Na 2 SO 3 aqueous solution the enhancement in the catalytic activity of the Zn(OH) 2 /Cd 0.3 Zn 0.7 S composite is due to the sulfurization of Zn(OH) 2 , while in the aqueous solution of ethanol the activation of the photocatalyst is associated with the formation of 3-Zn(OH) 2 . These results indicate great potential for Zn(OH) 2 /Cd 1Àx Zn x S photocatalysts, because a rather simple technique can be used for their synthesis, and for different sacricial agents the cocatalyst would transform to the desired form in situ during the process of hydrogen evolution and would provide a high photocatalytic activity.

Photocatalyst synthesis
The y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalysts with a Zn(OH) 2 weight content (y) from 10 to 30% were prepared via the precipitation of Zn(OH) 2 by sodium hydroxide from an aqueous solution of zinc nitrate Zn(NO 3 ) 2 , which contains suspended Cd 0.3 Zn 0.7 S nanoparticles, and was performed in accordance with the technique described in our previously published paper. 18 The y% ZnS/Cd 0.3 Zn 0.7 S photocatalysts with a ZnS weight content (y) from 0.5 to 10% were prepared via the precipitation of ZnS by sodium sulde from an aqueous solution of zinc nitrate Zn(NO 3 ) 2 , which contains suspended Cd 0.3 Zn 0.7 S nanoparticles. Also, b-Zn(OH) 2 was also prepared by a simple interaction of Zn(NO 3 ) 2 with NaOH followed by so drying for several days. Note that the hydrogen evolution from the ethanol solution was a slow process, and to increase the photocatalytic activity in these experiments, the composite photocatalysts were additionally modied with platinum (1 wt%). Pt deposition was performed via the chemical reduction of H 2 PtCl 6 with NaBH 4 according to the previously published technique. 18 The Pt-loaded photocatalysts are named in this paper as 1% Pt/y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S. All of the reagents used during the synthesis were of analytic grade and were used as supplied without further purication.

Photocatalyst characterization
The photocatalysts were characterized before and aer the catalytic tests using X-ray diffraction (XRD), UV-Vis diffuse reectance spectroscopy (UV-Vis DRS), and transmission electron microscopy (TEM) techniques. The XRD patterns were recorded on a D8 Advance diffractometer (Bruker, Germany) using Cu Ka radiation. The scanning was performed in the 2q range from 15 to 65 with a step of 0.05 and an acquisition time of 10 s at each point. The peak deconvolution was carried out using two Gaussian functions, and the composition of Cd 1Àx -Zn x S was determined from the angular positions of the functions. The content of the 3-Zn(OH) 2 phases was calculated with the TOPAS package (General Prole and Structure Analysis Soware for Powder Diffraction Data, Bruker AXS GmBH, Germany). The weight content of Cd 0.3 Zn 0.7 S was assumed to be a constant. The hydrozincite (Zn 5 (CO 3 ) 2 (OH) 6 ) and b-Zn(OH) 2 phases were not suitable for quantitative analysis with the TOPAS package owing to their imperfect structure and insufficient data. Therefore, their total content was estimated to be the difference between 100% and the calculated contents of Cd 1Àx Zn x S and 3-Zn(OH) 2. The UV-Vis diffuse reectance spectra of the photocatalysts were recorded at room temperature in the range of 250-850 nm with a resolution of 1 nm using a Cary 300 UV-Vis spectrophotometer from Agilent (USA) equipped with a DRA-30I diffuse reectance accessory. Special pre-packed polytetrauoroethylene (PTFE) from Agilent (USA) was used as a reectance standard. The high resolution TEM (HRTEM) images were obtained with a JEM-2010 transmission electron microscope (JEOL, Japan) operated at an accelerating voltage of 200 kV which provides a lattice resolution of 0.14 nm.
The local elemental composition was analyzed using a Phoenix energy-dispersive spectrometer (EDAX, USA) equipped with a Si(Li) detector with an energy resolution of 130 eV. X-ray photoelectron spectroscopy (XPS) measurements were performed using a SPECS photoelectron spectrometer (SPECS Surface Nano Analysis GmbH, Germany) equipped with an XR-50M X-ray source, a FOCUS-500 ellipsoidal crystal monochromator, and a PHOIBOS-150 hemispherical electron energy analyzer. The valence band and core-level spectra were obtained using monochromatic Al Ka radiation.

Photocatalytic activity measurement
The synthesized catalysts were studied for photocatalytic hydrogen evolution from an aqueous solution of inorganic or organic sacricial agents according to the technique described elsewhere. 18 In the case of the inorganic agent, 50 mg of y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalyst was suspended in 100 mL of 0.1 M Na 2 S/0.1 M Na 2 SO 3 aqueous solution, purged with argon for 30 min, and illuminated with a 450-LED (30 W, China). In the case of the organic donor, 50 mg of 1% Pt/y% Zn(OH) 2 / Cd 0.3 Zn 0.7 S was suspended in the solution, which contained 10 mL of ethanol, 90 mL of water, and 400 mg of NaOH. Then, the suspension was purged with argon for 30 min and illuminated with a 450-LED (30 W, China). The amount of evolved hydrogen was measured with a Khromos GCh-1000 gas chromatograph (Khromos, Russia) equipped with a zeolite column and a thermal conductivity detector. Argon was used as the carrier gas. The photocatalytic activity of the pristine Cd 0.3 Zn 0.7 S and synthesized composites was estimated as the rate of hydrogen evolution (mmol min À1 ). Note that the photocatalytic activity of the samples without any sacricial agents was very low (e.g. 0.008 AE 0.002 mmol min À1 for Cd 0.3 Zn 0.7 S).

Results and discussion
Co-catalysts in composite photocatalysts may provide a substantial increase in the rate of photocatalytic hydrogen evolution from aqueous solutions of various sacricial agents. The photocatalysts based on the composition of Cd 1Àx Zn x S and Zn(OH) 2 show a high activity in this process. Earlier, we studied the photocatalytic activity of the y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S and 1% Pt/ y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S composites for hydrogen production from aqueous solutions of Na 2 S/Na 2 SO 3 and ethanol, respectively. 18 The reaction rate increased as the Zn(OH) 2 content was increased up to 10-20 wt% owing to the formation of heterojunctions between the components of the photocatalyst. In the case of the Na 2 S/Na 2 SO 3 sacricial agent, the maximum reaction rate was observed for the 20% Zn(OH) 2 /Cd 0.3 Zn 0.7 S sample. In the case of the ethanol sacricial agent, a much lower rate of hydrogen production was observed, and the composites were additionally modied with platinum (1 wt%) to increase the photocatalytic activity. For ethanol as a sacricial agent, the maximum reaction rate was observed for the 1% Pt/10% Zn(OH) 2 /Cd 0.3 Zn 0.7 S sample. In addition to the fact that Zn(OH) 2 itself is not active in hydrogen production under visible light (Table 1), it may be considered as an active component of the composite photocatalyst. 18 The level of the conduction band (CB) for Zn(OH) 2 is close to À0.3 V versus normal hydrogen electrode (NHE), whereas for Cd 0.3 Zn 0.7 S, which is a light absorbing phase, the CB level is located at À0.6 V versus NHE. Therefore, the photogenerated electrons can transfer from the sulde to the hydroxide nanoparticles, 2 and the increased activity of the Zn(OH) 2 /Cd 0.3 Zn 0.7 S composites is due to the heterojunctions between the components of the photocatalyst and the improved separation of the charge carriers.
At the same time for both agents, a further increase in the content of Zn(OH) 2 to greater than 10-20 wt% led to a decrease in the hydrogen production rate. The band gaps of Cd 0.3 Zn 0.7 S and Zn(OH) 2 are $2.7 and 5.1 eV, respectively, and only Cd 0.3 Zn 0.7 S can be activated under visible light. Therefore, the low activity of the composite at a high content of Zn(OH) 2 is due to a decrease in the amount of the Cd 0.3 Zn 0.7 S phase, which can absorb visible light.
Based on the points mentioned above, the composite Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalysts with a content of Zn(OH) 2 from 10 to 30% were selected for the experiments and detailed investigations in this study. The y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S and 1% Pt/y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S composites were tested in the cyclic experiments for hydrogen production with two sacricial agents (Na 2 S/Na 2 SO 3 and ethanol) under the same conditions to compare the transformation mechanism. For both sacricial agents, the pristine Cd 0.3 Zn 0.7 S and 1% Pt/Cd 0.3 Zn 0.7 S samples lost their catalytic activity aer several runs of hydrogen production (Table 1). This behavior may be due to the selfoxidation of the sulde surface by the photogenerated holes. 20 In contrast, the strong activation of the composite Zn(OH) 2 / Cd 0.3 Zn 0.7 S photocatalysts was achieved during three catalytic runs for both the inorganic and organic sacricial agents (Table 1).
We have previously shown that the activation of 1% Pt/y% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalysts in the case of hydrogen evolution from aqueous ethanol solution is caused by the transformation of b-Zn(OH) 2 to 3-Zn(OH) 2 . 18,21 The linear correlation between the reaction rate and the content of 3-Zn(OH) 2 in the composite photocatalyst was found. 22 Based on these results, the same transformation may be suggested for the photocatalytic hydrogen evolution from the Na 2 S/Na 2 SO 3 aqueous solution. To check this suggestion, a detailed analysis of the photocatalyst composition during long-term hydrogen evolution was performed using several characterization methods. It should be noted that the 20% Zn(OH) 2 /Cd 0.3 Zn 0.7 S and 1% Pt/10% Zn(OH) 2 /Cd 0.3 Zn 0.7 S samples had the maximum activity for Na 2 S/Na 2 SO 3 and ethanol sacricial agents, respectively, but these samples were not suitable for correct identication using XRD analysis owing to the low content of Zn(OH) 2 . This was the reason for the selection of the 30% Zn(OH) 2 / Cd 0.3 Zn 0.7 S and 1% Pt/30% Zn(OH) 2 /Cd 0.3 Zn 0.7 S samples with a Zn(OH) 2 content of 30 wt% for detailed analysis of the phase transformation. To support the ndings and to extend them to the samples with a reduced content of Zn(OH) 2 , the 10% Zn(OH) 2 /Cd 0.3 Zn 0.7 S and 20% Zn(OH) 2 /Cd 0.3 Zn 0.7 S samples were also tested in four consecutive runs for hydrogen evolution from the Na 2 S/Na 2 SO 3 aqueous solution. However, only the phase compositions of the fresh samples and the samples aer the fourth photocatalytic run were analyzed. Fig. 1a shows the XRD patterns for the 30% Zn(OH) 2 / Cd 0.3 Zn 0.7 S photocatalyst before and aer the hydrogen evolution runs. Three broad peaks located at 25-30, 42-50, and 52-60 are observed for the fresh photocatalyst. These peaks can be attributed to the Cd 1Àx Zn x S solid solution of x $ 0.7. Narrow peaks for the b-Zn(OH) 2 and 3-Zn(OH) 2 phases are also observed. At the same time, no peaks attributed to Zn(OH) 2 can  (Fig. 1a). An amorphization of zinc hydroxide or a decrease in its content during the runs may be the reason for this. Additionally, aer the rst run, the broad peaks attributed to the Cd 1Àx Zn x S solid solution became asymmetrical owing to the appearance of a signal from the ZnS phase (Fig. 1b). The formation of ZnS may result from the sulfurization of Zn(OH) 2 in the Na 2 S/Na 2 SO 3 aqueous solution, because the area of the ZnS peak grew aer each run of hydrogen evolution (Fig. 1b). The same behavior was observed for the 10% Zn(OH) 2 / Cd 0.3 Zn 0.7 S and 20% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalysts aer the fourth run (Fig. S1 in the ESI †). This statement also supported the experiments on Zn(OH) 2 alone, aer four runs in the Na 2 S/Na 2 SO 3 aqueous solution the pristine Zn(OH) 2 underwent partial sulfurization (see Fig. S2 in ESI †). Unfortunately, we cannot exactly calculate the ratio of Cd 0.3 Zn 0.7 S to ZnS, because both phases have a disordered structure. The TEM and HRTEM images of the 30% Zn(OH) 2 / Cd 0.3 Zn 0.7 S before and aer two irradiation runs were used to evaluate the degree of transformation of Zn(OH) 2 to ZnS ( Fig. 2   and 3). Fig. 2 shows that the fresh 30% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalyst consists of two different phases, namely, the Cd 0.3 Zn 0.7 S solid solution (d 100 ¼ 0.32 nm) and Zn(OH) 2 (d 200 ¼ 0.32 nm). Fig. 3 shows that the same phases are presented in the photocatalyst aer two runs of hydrogen evolution. In addition to these phases, the lattice fringe of 0.30 nm can be attributed to the (002) plane of ZnS. It indicates that zinc hydroxide is present in the sample, but the size of the Zn(OH) 2 particle is much smaller.
The lattice fringes corresponding to Cd 1Àx Zn x S and ZnS are close to each other, and to conrm the presence of the ZnS phase the samples were additionally investigated using the EDX technique (Fig. 4). The lines of Cd, Zn, and S are observed in the EDX spectrum of the fresh 30% Zn(OH) 2 /Cd 0.3 Zn 0.7 S sample, while the line for O is very low (Fig. 4a). These data are attributed to the solid solution of Cd 1Àx Zn x S. Fig. 4b shows the EDX spectrum for the domains depicted in Fig. 2 (right) and conrms the presence of zinc hydroxide only. The lines, which can be attributed to Zn and S (Fig. 4c), as well as to Zn, Cd, and S (Fig. 4d), were observed for the sample aer two runs of hydrogen evolution. In both cases, the content of oxygen was  very low. Therefore, the area in Fig. 4c can be attributed to the ZnS phase, while the area in Fig. 4d corresponds to Cd 1Àx Zn x S. The data from the EDX analysis for the studied samples are summarized in Table 2. They conrm the presence of two types of domains, as mentioned above. It should be noted that the O content for the fresh photocatalyst varied from 9 to 54 at%, while for the sample aer two runs of hydrogen evolution this value was in the range of 3-33 at%. Simultaneously, the S content increased during the long-term photocatalytic hydrogen production. The lowest S content was equal to 2 at% for the fresh photocatalyst and 25 at% for the photocatalyst aer long-term hydrogen production. The highest S content  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 1341-1350 | 1345 also grew from 42 to 52 at%. This increase in the S content with a simultaneous decrease in the O content conrms the partial sulfurization of zinc hydroxide during the photocatalytic hydrogen production. Fig. 5a shows the diffuse reectance spectra of the fresh photocatalyst and the sample aer the rst and fourth runs of hydrogen evolution. The fresh photocatalyst had a spectrum that is typical of materials based on a Cd 0.3 Zn 0.7 S solid solution. 4 No peaks that are typical of Zn(OH) 2 were observed for this photocatalyst. The spectra of the composite samples aer the rst and fourth runs were similar, which corresponds to the light absorption by the Cd 0.3 Zn 0.7 S phase (Fig. 4). According to the spectra for individual phases (Fig. 5b), an additional shoulder at $340 nm in these spectra can be attributed to the ZnS phase. 23 The DRS data conrm the sulfurization of the cocatalyst and completely agrees with the data from other methods.
Therefore, the sulfurization of the Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalysts occurs during the photocatalytic hydrogen production from the Na 2 S/Na 2 SO 3 aqueous solution under visible light. The XRD patterns of the 30% Zn(OH) 2 /Cd 0.3 Zn 0.7 S sample aer the second to the fourth photocatalytic runs in Fig. 1b are quite similar and conrm the presence of ZnS and Cd 0.3 Zn 0.7 S only. Also, Table 1 shows that the rate of hydrogen evolution almost doubles aer the rst run, and then it does not change signicantly. Based on these results, we can conclude that the transformation of b-Zn(OH) 2 to ZnS is quite a fast process. In addition to the fact that zinc hydroxide nanoparticles are present in the sample (Fig. 3), their amount is probably small. ZnS is known to have a weak absorption in the visible region, 23 and both Cd 0.3 Zn 0.7 S and ZnS can be activated under visible light. However, we believe that the heterojunctions between ZnS and Cd 0.3 Zn 0.7 S are probably realized in the multiphase sample and are responsible for the improvement of the photocatalytic activity. The kinetic data (Table 1) conrms this statement, because the hydrogen production rate for the 20 or 30% Zn(OH) 2 / Cd 0.3 Zn 0.7 S samples, which transforms to the form of ZnS/ Cd 0.3 Zn 0.7 S during the reaction, is higher than the rate for Cd 0.3 Zn 0.7 S (3.2 mmol min À1 ) or ZnS (0.3 mmol min À1 ).
To conrm the high activity of the ZnS/Cd 1Àx Zn x S heterostructures, the photocatalytic activity of the y% ZnS/Cd 0.3 Zn 0.7 S samples, prepared via a simple deposition of ZnS onto the surface of Cd 0.3 Zn 0.7 S, was also studied and is listed in Table 1. The reaction rate slightly increased as the ZnS content was increased up to 1 wt%. The further increase in the ZnS content led to a decrease in the reaction rate (Table 1). This observation conrms that the formation of ZnS from Zn(OH) 2 in situ during the process of hydrogen production is benecial to a high photocatalytic activity compared to the direct synthesis of the ZnS/Cd 0.3 Zn 0.7 S composites. This is probably due to the closer contact between the formed nanoparticles in the rst case.  In addition to Na 2 S/Na 2 SO 3 , we also checked the data for ethanol as a sacricial agent, and the transformations of the 1% Pt/30% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalyst under irradiation in an aqueous solution of ethanol were studied in detail. Fig. 6 shows the XRD patterns of 1% Pt/30% Zn(OH) 2 /Cd 0.3 Zn 0.7 S before and aer the second to the fourth runs of hydrogen evolution. Cd 0.3 Zn 0.7 S and a mixture of the b-Zn(OH) 2 and 3-Zn(OH) 2 phases were detected in the composition of the fresh photocatalyst. During the photocatalytic hydrogen production, the intensity of the peaks, which correspond to b-Zn(OH) 2 , decreased, whereas the intensity of the 3-Zn(OH) 2 peaks increased in contrast, and the reaction rate also increased (see Fig. 7 and Table 1). Quantitative analysis of the XRD data conrms this observation (see Table 3). 11 wt% b-Zn(OH) 2 and 11 wt% 3-Zn(OH) 2 was detected in the fresh photocatalyst, whereas only the 3-Zn(OH) 2 phase (22 wt%) was detected in the photocatalyst aer four runs of hydrogen evolution. Additionally, the hydrozincite-like phase Zn 5 (CO 3 ) 2Àx (OH) 6+2x $yH 2 O, which is referred to as HZ, was detected aer the rst run, but it completely disappeared at the end of the fourth run. The HZ phase may be formed during the wet stage of the synthesis owing to the presence of dissolved CO 2 in solution and/or during drying of the samples in air. 24 We have previously shown that the presence of the 3-Zn(OH) 2 phase is benecial for H 2 production, whereas both b-Zn(OH) 2 and HZ are inactive in this process. 22 As further support for this statement, the reaction rate of the hydrogen production in this study monotonically increased as the content of 3-Zn(OH) 2 increased (see Fig. 7).
The most active photocatalyst, 1% Pt/30% Zn(OH) 2 / Cd 0.3 Zn 0.7 S, was also characterized using an XPS technique before and aer four runs of hydrogen evolution (Table 4). Aer the long-term photocatalytic process, the surface ratio of   cadmium to zinc fell from 3.9 to 1.8. At the same time, the surface ratio of sulfur to oxygen increased in contrast. Therefore, b-Zn(OH) 2 , which according to the synthesis technique probably covers the surface of Cd 0.3 Zn 0.7 S, transforms to 3-Zn(OH) 2 with a high crystalline size and releases from the surface of the Cd 0.3 Zn 0.7 S sulde. The surface ratio and oxidation state of the deposited platinum (Pt 0 /Pt 2+ ) were similar before and aer four runs (Table 4). This result indicates that the change in activity of the composite photocatalysts is caused by bulk transformations rather than the surface transformations. Therefore, the transformation of 1% Pt/b-Zn(OH) 2 / Cd 0.3 Zn 0.7 S to 1% Pt/3-Zn(OH) 2 /Cd 0.3 Zn 0.7 S leads to a signicant growth in the H 2 production rate from the ethanol aqueous solution. It is important to note that up to a seven-fold increase in the rate of the photocatalytic hydrogen evolution was achieved in the case of the ethanol donor during the four catalytic runs, whereas for the hydrogen evolution from the Na 2 S/Na 2 SO 3 aqueous solution the rate increased only twice (Table 1). Therefore, the heterojunctions between 3-Zn(OH) 2 and Cd 0.3 Zn 0.7 S can be concluded to be more efficient than those between ZnS and Cd 0.3 Zn 0.7 S.
To summarize, the zinc hydroxide co-catalyst in the Zn(OH) 2 / Cd 0.3 Zn 0.7 S composite photocatalysts undergoes sulfurization during the photocatalytic hydrogen production from Na 2 S/ Na 2 SO 3 solution, whereas during the hydrogen evolution from ethanol aqueous solution the transformation of the b-Zn(OH) 2 co-catalyst to 3-Zn(OH) 2 was realized. A signicant increase in the activity accompanies both types of transformations. The highest hydrogen production rates are 3200 mmol g À1 h À1 for 1% Pt/10% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalyst (ethanol) and 8900 mmol g À1 h À1 for the 20% Zn(OH) 2 /Cd 0.3 Zn 0.7 S photocatalyst (Na 2 S/Na 2 SO 3 ) which are comparable with recently published data (see Table 5). [25][26][27][28][29][30][31][32][33][34][35] Therefore, the sacricial agent as a reaction media plays a key role in the photocatalyst transformation and affects its catalytic activity (see graphical abstract). This is a very promising result, because it is possible to synthesize a composite photocatalyst by a rather simple technique, and the co-catalyst could be tuned in situ during the process of photocatalytic hydrogen evolution. In our previous works, we made a signicant efforts to obtain a system with zinc hydroxide for the desired 3-modication. 21 The present study shows that the initial state of the photocatalyst is not as important for its activity.

Conclusions
In this study, we have shown for the rst time that the nature of the sacricial agent strongly affects the transformation of the composite photocatalyst during photocatalytic hydrogen production. The composites, which consist of a Cd 0.3 Zn 0.7 S photocatalyst and a Zn(OH) 2 co-catalyst, are active during the  photocatalytic hydrogen evolution under visible light with both inorganic (Na 2 S/Na 2 SO 3 ) and organic (ethanol) sacricial agents, but in Na 2 S/Na 2 SO 3 solution, the Zn(OH) 2 co-catalyst transforms to zinc sulde, whereas in ethanol solution, b-Zn(OH) 2 transforms to 3-Zn(OH) 2 . Both transformations are accompanied by a signicant increase in the photocatalytic activity that is probably due to heterojunctions in the formed pairs of ZnS/Cd 0.3 Zn 0.7 S and 3-Zn(OH) 2 /Cd 0.3 Zn 0.7 S, respectively.

Conflicts of interest
There are no conicts of interest to declare.