Facile Synthesis of Novel CaIn2S4/ZnIn2S4 Composites with Efficient Performance for Photocatalytic Reduction of Cr(VI) under Simulated Sunlight Irradiation

A series of novel and efficient heterostructured composites CaIn2S4/ZnIn2S4 have been synthesized using a facile hydrothermal method. XRD patterns indicate the as-prepared catalysts are two-phase composites of cubic phase CaIn2S4 and hexagonal phase ZnIn2S4. FESEM (field emission scanning electron microscope) images display that the synthesized composites are composed of flower-like microspheres with wide diameter distribution. UV–Vis diffuse reflectance spectra (DRS) show that the optical absorption edges of the CaIn2S4/ZnIn2S4 composites shift toward longer wavelengths with the increase of the CaIn2S4 component. The photocatalytic activities of the as-synthesized composites are investigated by using the aqueous-phase Cr(VI) reduction under simulated sunlight irradiation. This is the first report on the application of the CaIn2S4/ZnIn2S4 composites as stable and efficient photocatalysts for the Cr(VI) reduction. The fabricated CaIn2S4/ZnIn2S4 composites possess higher photocatalytic performance in comparison with pristine CaIn2S4 or ZnIn2S4. The CaIn2S4/ZnIn2S4 composite with a CaIn2S4 molar content of 30% exhibits the optimum photocatalytic activity. The primary reason for the significantly enhanced photoreduction activity is proved to be the substantially improved separation efficiency of photogenerated electrons/holes caused by forming the CaIn2S4/ZnIn2S4 heterostructured composites. The efficient charge separation can be evidenced by steady-state photoluminescence spectra (PLs) and transient photocurrent response. Based on the charge transfer between CaIn2S4 and ZnIn2S4, an enhancement mechanism of photocatalytic activity and stability for the Cr(VI) reduction is proposed.


Introduction
Semiconductor-based photocatalytic technology has exhibited great potential in green controlling of environmental contaminants and converting CO 2 to valuable chemicals [1][2][3][4][5][6]. However, the most studied photocatalysts (such as TiO 2 , ZnO, and ZnS) are active only under ultraviolet light irradiation. In view of the practical applications, visible light active photocatalysts with narrow bandgaps are greatly desirable. A great deal of research and efforts have been devoted to fabricating and synthesizing energy conversion, photocatalytic reduction is postulated to be an efficient and green technology for the elimination of Cr(VI) ions from contaminated water. The highly toxic Cr(VI) can be photoreduced to less harmful Cr(III) by means of a certain photocatalyst and reaction system.
In this study, flower-like CaIn 2 S 4 /ZnIn 2 S 4 heterojunction composites are successfully synthesized by using a one-step hydrothermal method, and the photocatalytic performance of the Cr(VI) reduction is investigated under simulated sunlight illumination. To our knowledge, this is the first report about the CaIn 2 S 4 /ZnIn 2 S 4 heterojunction composites for the photocatalytic reduction of Cr(VI). The heterojunction composites exhibit much higher reduction efficiency of Cr(VI) than the pure CaIn 2 S 4 or ZnIn 2 S 4 . Meanwhile, the CaIn 2 S 4 /ZnIn 2 S 4 composites present excellent solar stability, and the photocorrosion of ZnIn 2 S 4 is dramatically inhibited by constructing the CaIn 2 S 4 /ZnIn 2 S 4 composites. The detailed mechanism of enhanced photocatalytic performance for the Cr(VI) reduction over the CaIn 2 S 4 /ZnIn 2 S 4 composites is also proposed.
For comparison, the pure ZnIn 2 S 4 was prepared by using a similar process without Ca(NO 3 ) 3 . 1.0 mmol ZnSO 4 ·7H 2 O was put into 40 mL of distilled water and stirred for 30 min. Then, 2 mmol InCl 3 ·4H 2 O and 8 mmol TAA were added to the above-obtained solution and stirred for another 60 min. This mixture was subsequently transferred into a Teflon-lined steel autoclave and heated at 180 • C for 12 h. The precipitate was finally washed with deionized water and dried at 60 • C for 8 h to obtain pure ZnIn 2 S 4 . For the preparation of the pure CaIn 2 S 4 , 1.0 mmol Ca(NO 3 ) 3 ·4H 2 O was added into 40 mL of distilled water and stirred for 30 min. After that, 2 mmol InCl 3 ·4H 2 O and 8 mmol TAA were put into the above solution and then stirred for another 60 min. The obtained mixture was transferred into a Teflon-lined steel autoclave and heated at 180 • C for 12 h. The orange precipitate was finally washed with deionized water and dried at 60 • C for 8 h to obtain pure CaIn 2 S 4 .

Material Characterization
The crystal structures of the as-prepared composites were investigated by using an X-ray diffractometer (XRD, Bruker D8 Advance, Karlsruhe, Germany), with Cu Kα radiation (λ = 0.15405 nm). The morphologies of the synthesized samples were observed by a field emission scanning electron microscope (FEI, Quanta250, Hillsboro, OR, USA). The X-ray photoelectron Nanomaterials 2018, 8,472 4 of 17 spectroscopy (XPS) experiment was carried out by using a Thermo Scientific ESCALAB 250xi system (Waltham, MA, USA), equipped with an Al anode. The BET (Brunauer-Emmett-Teller) surface areas of the composite samples were determined on a surface area analyzer (Autosorb-IQ, Quantachrome, Boynton Beach, FL, USA). UV-visible diffuse reflectance spectra (DRS) were collected on a Shimadzu UV-2550 spectrometer (Kyoto, Japan) using BaSO 4 as the reflectance standard. Photoluminescence spectra (PLs) were measured by using an Edinburgh FLS 980 fluorescence spectrometer (Livingston, UK), with a 330 nm excitation wavelength. The transient photocurrent measurement was performed on a CHI electrochemical workstation (CHI 760D, Shanghai, China) in the standard three-electrode system. Ag/AgCl electrode and a platinum wire were employed as the reference electrode and the counter electrode, respectively. The working electrodes were prepared according to our previous report [32]. The photocurrent was measured in the electrolyte of a 0.5 mol/L Na 2 SO 4 aqueous solution (pH~6.8) at a bias of +0.5 V. The light source was the same as that used in the photocatalytic experiments.

Photocatalytic Reduction of Cr(VI)
For the photocatalytic reduction of Cr(VI) under simulated sunlight irradiation, a 300 W xenon lamp (PLS-SXE 300, Perfect Light Co. Ltd., Beijing, China) was used as the light source. The incident light intensity was 75 mW/cm 2 . In the typical photocatalytic test, 50 mg of photocatalyst was suspended in 50 mL of 20 mg/L Cr(VI) aqueous solution. After adding 5 mg of ammonium oxalate (a scavenger for photo-hole), the suspension was magnetically stirred in the dark for 30 min to establish an adsorption-desorption equilibrium. As the photocatalytic reduction proceeded, 3 mL of the reaction solution was taken out at a given time interval and centrifuged to remove the catalyst particles. The concentrations of Cr(VI) were determined colorimetrically at 540 nm using the diphenylcarbazide (DPC) method on a Shimadzu UV-160A UV-Vis spectrophotometer [33]. In addition, the total Cr ions concentrations were measured by an inductively coupled plasma-optical emission spectrophotometer (Agilent 725 ICP-OES, Palo Alto, CA, USA).
The apparent quantum efficiency (AQE) of Cr(VI) the photocatalytic reduction was measured under the same photocatalytic reaction conditions, except by using a 420 nm bandpass filter. The apparent quantum efficiencies were calculated according to the following equation: To evaluate the catalytic stability of the CaIn 2 S 4 /ZnIn 2 S 4 composites, the photocatalysts-after the first run for the Cr(VI) reduction-were separated by centrifugation from the suspension and washed with 1 M nitrite acid solution and deionized water. After being dried at 60 • C, the recovered photocatalysts were reused for the next run of the photocatalytic Cr(VI) reductions under the same experimental conditions.

XRD Analysis and BET Surface Area
The phase composition and crystalline properties of the CaIn 2 S 4 /ZnIn 2 S 4 composite samples were analyzed by XRD. Figure 1 displays the XRD patterns of the pure ZnIn 2 S 4 , CaIn 2 S 4 , and CaIn 2 S 4 /ZnIn 2 S 4 composites. As indicated in Figure 1, the diffraction peaks of the pure ZnIn 2 S 4 at 2θ = 21.  [24]. No other diffraction peaks were detected in the XRD pattern of ZnIn 2 S 4 , indicating that the obtained ZnIn 2 S 4 is highly pure. With the addition of Ca(NO 3 ) 3 in the preparation process, new diffraction peaks appeared in the XRD patterns, which belonged to the characteristic peaks of the cubic phase CaIn 2 S 4 (JCPDS No. 16-0341) [25][26][27]. It can be also observed from Figure 1 that the intensity of the diffraction peaks belonging to the cubic CaIn 2 S 4 phase increased gradually by increasing the mole proportion of CaIn 2 S 4 , which indicated the existence of both CaIn 2 S 4 and ZnIn 2 S 4 in the as-synthesized composites. In addition, the diffraction peaks corresponding to binary sulfides, oxides, and other new compounds were not observed, indicating that CaIn 2 S 4 and ZnIn 2 S 4 maintained the pure phase and no impurities were formed in the obtained CaIn 2 S 4 /ZnIn 2 S 4 composites.  [25][26][27]. It can be also observed from Figure 1 that the intensity of the diffraction peaks belonging to the cubic CaIn2S4 phase increased gradually by increasing the mole proportion of CaIn2S4, which indicated the existence of both CaIn2S4 and ZnIn2S4 in the as-synthesized composites. In addition, the diffraction peaks corresponding to binary sulfides, oxides, and other new compounds were not observed, indicating that CaIn2S4 and ZnIn2S4 maintained the pure phase and no impurities were formed in the obtained CaIn2S4/ZnIn2S4 composites. The BET surface area of the pure ZnIn2S4, CaIn2S4, and CaIn2S4/ZnIn2S4 composites was measured, and the results are summarized in Table 1. As can be noted from Table 1, the specific surface area of the CaIn2S4/ZnIn2S4 composites decreased slightly with the continuous increment of CaIn2S4 component. The actual molar ratios of Ca-Zn in the CaIn2S4/ZnIn2S4 composites were determined by inductively coupled plasma elemental analysis, as presented in Table 1. It can be observed from Table 1 that the experimentally measured molar ratios of Ca-Zn in the synthesized composites were close to those of the added proportions in the preparation process.

SEM and Elemental Mapping Analysis
The SEM images of the as-synthesized pure ZnIn2S4, CaIn2S4, and CaIn2S4/ZnIn2S4 composites are indicated in Figure 2. As can be observed from Figure 2, the pure ZnIn2S4 was composed of hierarchical microspheres with a wide distribution of diameter, which was consistent with the  The BET surface area of the pure ZnIn 2 S 4 , CaIn 2 S 4 , and CaIn 2 S 4 /ZnIn 2 S 4 composites was measured, and the results are summarized in Table 1. As can be noted from Table 1, the specific surface area of the CaIn 2 S 4 /ZnIn 2 S 4 composites decreased slightly with the continuous increment of CaIn 2 S 4 component. The actual molar ratios of Ca-Zn in the CaIn 2 S 4 /ZnIn 2 S 4 composites were determined by inductively coupled plasma elemental analysis, as presented in Table 1. It can be observed from Table 1 that the experimentally measured molar ratios of Ca-Zn in the synthesized composites were close to those of the added proportions in the preparation process.

SEM and Elemental Mapping Analysis
The SEM images of the as-synthesized pure ZnIn 2 S 4 , CaIn 2 S 4 , and CaIn 2 S 4 /ZnIn 2 S 4 composites are indicated in Figure 2. As can be observed from Figure 2, the pure ZnIn 2 S 4 was composed of hierarchical microspheres with a wide distribution of diameter, which was consistent with the previous reports [17,24]. Introducing the Ca component had almost no influence on the morphologies of the CaIn 2 S 4 /ZnIn 2 S 4 composites, which also exhibited flower-like microspheres constructed by numerous nanosheets in the form of random self-assembly. As a consequence, the porous structures with wide pore-size distribution could be expected. This would benefit the photocatalytic reaction by increasing the specific surface area. Moreover, the formation of the CaIn 2 S 4 /ZnIn 2 S 4 heterojunction was evidenced by the elemental mapping of the as-synthesized composites ( Figure 3). Maps of Zn-K, Ca-K, In-L, and S-K display the same shape and location, indicating the coexistence of ZnIn 2 S 4 and CaIn 2 S 4 components in the obtained composites. This provided solid evidence for the formation of CaIn 2 S 4 /ZnIn 2 S 4 heterostructured composites. previous reports [17,24]. Introducing the Ca component had almost no influence on the morphologies of the CaIn2S4/ZnIn2S4 composites, which also exhibited flower-like microspheres constructed by numerous nanosheets in the form of random self-assembly. As a consequence, the porous structures with wide pore-size distribution could be expected. This would benefit the photocatalytic reaction by increasing the specific surface area. Moreover, the formation of the CaIn2S4/ZnIn2S4 heterojunction was evidenced by the elemental mapping of the as-synthesized composites ( Figure 3). Maps of Zn-K, Ca-K, In-L, and S-K display the same shape and location, indicating the coexistence of ZnIn2S4 and CaIn2S4 components in the obtained composites. This provided solid evidence for the formation of CaIn2S4/ZnIn2S4 heterostructured composites.

XPS Analysis and Optical Properties
To determine the elemental composition and the corresponding chemical states of the synthesized composites, the XPS spectra of the 30% CaIn 2 S 4 /ZnIn 2 S 4 sample are indicated in Figure 4. As can be seen from Figure 4a, the survey spectrum indicated the presence of Zn, Ca, In, and S elements in the 30% CaIn 2 S 4 /ZnIn 2 S 4 sample. The high-resolution XPS spectrum for Zn is presented in Figure 4b. Two characteristic XPS signals were observed at binding energies of 1022.4 and 1045.3 eV, which were ascribed to Zn 2+ 2p 1/2 and Zn 2+ 2p 3/2 , respectively [33,34]. The high-resolution XPS spectrum of In 3d is displayed in Figure 4c. The characteristic peaks centered at 444.7 and 452.3 eV can be attributed to the In 3d 3/2 and In 3d 5/2 signals of In 3+ species, respectively [27,34]. In Figure 4d, the two peaks at the binding energies of 351.1 and 347.6 eV corresponded to the 2p 3/2 and 2p 1/2 levels of Ca 2+ . Figure 4e shows an XPS signal centered at 162.5 eV, which can be assigned to the 2p 1/2 level of S 2− in the as-prepared CaIn 2 S 4 /ZnIn 2 S 4 composites [34]. These results further indicate that CaIn 2 S 4 /ZnIn 2 S 4 composites can be successfully synthesized through a one-step hydrothermal reaction process.

XPS Analysis and Optical Properties
To determine the elemental composition and the corresponding chemical states of the synthesized composites, the XPS spectra of the 30% CaIn2S4/ZnIn2S4 sample are indicated in Figure  4. As can be seen from Figure 4a, the survey spectrum indicated the presence of Zn, Ca, In, and S elements in the 30% CaIn2S4/ZnIn2S4 sample. The high-resolution XPS spectrum for Zn is presented in Figure 4b. Two characteristic XPS signals were observed at binding energies of 1022.4 and 1045.3 eV, which were ascribed to Zn 2+ 2p1/2 and Zn 2+ 2p3/2, respectively [33,34]. The high-resolution XPS spectrum of In 3d is displayed in Figure 4c. The characteristic peaks centered at 444.7 and 452.3 eV can be attributed to the In 3d3/2 and In 3d5/2 signals of In 3+ species, respectively [27,34]. In Figure 4d, the two peaks at the binding energies of 351.1 and 347.6 eV corresponded to the 2p3/2 and 2p1/2 levels of Ca 2+ . Figure 4e shows an XPS signal centered at 162.5 eV, which can be assigned to the 2p1/2 level of S 2− in the as-prepared CaIn2S4/ZnIn2S4 composites [34]. These results further indicate that CaIn2S4/ZnIn2S4 composites can be successfully synthesized through a one-step hydrothermal reaction process.   The optical absorption properties of the as-obtained CaIn2S4/ZnIn2S4 composites were analyzed by UV-Vis DRS, and the results are depicted in Figure 5a. The absorption edges of the pure ZnIn2S4 and CaIn2S4 samples were at around 537 nm and 638 nm, respectively. As can be also noted from Figure 5a, the absorption edges of the CaIn2S4/ZnIn2S4 samples were gradually red-shifted from 537 to 570 nm as the molar percentage of the CaIn2S4 component increased to 50%. The photoresponse of the CaIn2S4/ZnIn2S4 composites in visible-light region was significantly improved by comparison with that of the pure ZnIn2S4. Moreover, the UV-Vis DRS shown in Figure 5a were all very steep, demonstrating the visible-light absorption was ascribed to the intrinsic band transition instead of the transition from impurity levels [35]. Based on the optical absorption theory of the bandgap semiconductor, the bandgap energy of the pure CaIn2S4 and ZnIn2S4 can be calculated by the Equation (1) [36] as follows: where α, v, Eg, and A represent the absorption coefficient, light frequency, bandgap energy, and a constant, respectively. The value of n depends on the type of optical transition of a semiconductor (n = 1 for the direct transition and n = 4 for the indirect transition). According to the previous literature [27,37], ZnIn2S4 and CaIn2S4 are direct-transition semiconductors, and thus, the bandgap energies of CaIn2S4 and ZnIn2S4 can be estimated from the plots of (αhv) 2 versus light energy (hv). As illustrated in Figure 5b, the estimated bandgaps were 2.03 and 2.43 eV for the pure CaIn2S4 and ZnIn2S4, respectively, which matched well with the literature values [15,27]. For the composite catalysts, the photocatalytic performance was primarily determined by the valence band (VB) and conduction band (CB) energy levels of the constituent semiconductors. Based on the following equations, the VB and CB positions of the pristine ZnIn2S4 and CaIn2S4 can be obtained: where EVB and ECB are the potential of the VB and CB edge, Eg is the bandgap energy, χ is the geometric mean of the absolute electronegativity of the constituent atoms in the semiconductor, and Ee is the energy of free electrons on the hydrogen scale with a value of 4.5 eV.  The optical absorption properties of the as-obtained CaIn 2 S 4 /ZnIn 2 S 4 composites were analyzed by UV-Vis DRS, and the results are depicted in Figure 5a. The absorption edges of the pure ZnIn 2 S 4 and CaIn 2 S 4 samples were at around 537 nm and 638 nm, respectively. As can be also noted from Figure 5a, the absorption edges of the CaIn 2 S 4 /ZnIn 2 S 4 samples were gradually red-shifted from 537 to 570 nm as the molar percentage of the CaIn 2 S 4 component increased to 50%. The photoresponse of the CaIn 2 S 4 /ZnIn 2 S 4 composites in visible-light region was significantly improved by comparison with that of the pure ZnIn 2 S 4 . Moreover, the UV-Vis DRS shown in Figure 5a were all very steep, demonstrating the visible-light absorption was ascribed to the intrinsic band transition instead of the transition from impurity levels [35]. Based on the optical absorption theory of the bandgap semiconductor, the bandgap energy of the pure CaIn 2 S 4 and ZnIn 2 S 4 can be calculated by the Equation (1) [36] as follows: where α, v, E g , and A represent the absorption coefficient, light frequency, bandgap energy, and a constant, respectively. The value of n depends on the type of optical transition of a semiconductor (n = 1 for the direct transition and n = 4 for the indirect transition). According to the previous literature [27,37], ZnIn 2 S 4 and CaIn 2 S 4 are direct-transition semiconductors, and thus, the bandgap energies of CaIn 2 S 4 and ZnIn 2 S 4 can be estimated from the plots of (αhv) 2 versus light energy (hv). As illustrated in Figure 5b, the estimated bandgaps were 2.03 and 2.43 eV for the pure CaIn 2 S 4 and ZnIn 2 S 4 , respectively, which matched well with the literature values [15,27]. For the composite catalysts, the photocatalytic performance was primarily determined by the valence band (VB) and conduction band (CB) energy levels of the constituent semiconductors. Based on the following equations, the VB and CB positions of the pristine ZnIn 2 S 4 and CaIn 2 S 4 can be obtained: where E VB and E CB are the potential of the VB and CB edge, E g is the bandgap energy, χ is the geometric mean of the absolute electronegativity of the constituent atoms in the semiconductor, and E e is the energy of free electrons on the hydrogen scale with a value of 4.5 eV.

Photocatalytic Activity
The photocatalytic performances of the ZnIn2S4, CaIn2S4, and CaIn2S4/ZnIn2S4 composites were investigated by the aqueous-phase Cr(VI) reduction under simulated sunlight irradiation ( Figure  6A). Cr(VI) cannot be reduced in the absence of light illumination or photocatalysts. The pure ZnIn2S4 displayed a relatively higher photocatalytic activity of the Cr(VI) reduction than the CaIn2S4 sample. About 42% of Cr(VI) was reduced over the pure CaIn2S4, while ZnIn2S4 can reduce 63% of Cr(VI) after irradiation of 30 min. All the CaIn2S4/ZnIn2S4 composites exhibited higher photocatalytic efficiency than the pristine CaIn2S4 and ZnIn2S4, indicating that the combination of CaIn2S4 and ZnIn2S4 can improve the photocatalytic reduction performance of these composites.

Photocatalytic Activity
The photocatalytic performances of the ZnIn 2 S 4 , CaIn 2 S 4 , and CaIn 2 S 4 /ZnIn 2 S 4 composites were investigated by the aqueous-phase Cr(VI) reduction under simulated sunlight irradiation ( Figure 6A). Cr(VI) cannot be reduced in the absence of light illumination or photocatalysts. The pure ZnIn 2 S 4 displayed a relatively higher photocatalytic activity of the Cr(VI) reduction than the CaIn 2 S 4 sample. About 42% of Cr(VI) was reduced over the pure CaIn 2 S 4 , while ZnIn 2 S 4 can reduce 63% of Cr(VI) after irradiation of 30 min. All the CaIn 2 S 4 /ZnIn 2 S 4 composites exhibited higher photocatalytic efficiency than the pristine CaIn 2 S 4 and ZnIn 2 S 4 , indicating that the combination of CaIn 2 S 4 and ZnIn 2 S 4 can improve the photocatalytic reduction performance of these composites.

Photocatalytic Activity
The photocatalytic performances of the ZnIn2S4, CaIn2S4, and CaIn2S4/ZnIn2S4 composites were investigated by the aqueous-phase Cr(VI) reduction under simulated sunlight irradiation ( Figure  6A). Cr(VI) cannot be reduced in the absence of light illumination or photocatalysts. The pure ZnIn2S4 displayed a relatively higher photocatalytic activity of the Cr(VI) reduction than the CaIn2S4 sample. About 42% of Cr(VI) was reduced over the pure CaIn2S4, while ZnIn2S4 can reduce 63% of Cr(VI) after irradiation of 30 min. All the CaIn2S4/ZnIn2S4 composites exhibited higher photocatalytic efficiency than the pristine CaIn2S4 and ZnIn2S4, indicating that the combination of CaIn2S4 and ZnIn2S4 can improve the photocatalytic reduction performance of these composites. For the synthesized CaIn2S4/ZnIn2S4 composite catalysts, the photocatalytic were closely associated with the contents of the CaIn2S4 component. As can be noted from Figure 6A, the photocatalytic activities of the CaIn2S4/ZnIn2S4 composite catalysts increased with the increment of component CaIn2S4. The 30% CaIn2S4/ZnIn2S4 composite photocatalyst exhibited the highest activity for the Cr(VI) reduction, whereas the greater increase in the amount of CaIn2S4 resulted in a decrease in the Cr(VI) reduction rates. It can be ascribed to the low photocatalytic activity of the pure CaIn2S4 because of the slow separation of the photogenerated charge carriers. Moreover, the apparent quantum efficiencies (AQE) of the Cr(VI) photocatalytic reduction over the synthesized composites were also calculated, and the corresponding results are summarized in Table 1. As indicated in Table 1, the AQE of the 30% CaIn2S4/ZnIn2S4 nanocomposite catalyst reached 6.6%, which presented higher than that of the pure ZnIn2S4 (3.7%) or pure CaIn2S4 (2.3%). In addition, the AQE for the Cr(VI) reduction increased gradually with the increase of the CaIn2S4 constituent when the addition ratio of the CaIn2S4 precursor was no more than 30%. However, further increasing the molar ratio of CaIn2S4 led to a decrease in the AQE. This suggests that the heterostructured composites containing a suitable amount of CaIn2S4 and ZnIn2S4 contributed to improving optimally the photoactivity for the Cr(VI) reduction. Therefore, the optimal 30% CaIn2S4/ZnIn2S4 has been chosen as a representative catalyst for the following studies. For the synthesized CaIn 2 S 4 /ZnIn 2 S 4 composite catalysts, the photocatalytic were closely associated with the contents of the CaIn 2 S 4 component. As can be noted from Figure 6A, the photocatalytic activities of the CaIn 2 S 4 /ZnIn 2 S 4 composite catalysts increased with the increment of component CaIn 2 S 4 . The 30% CaIn 2 S 4 /ZnIn 2 S 4 composite photocatalyst exhibited the highest activity for the Cr(VI) reduction, whereas the greater increase in the amount of CaIn 2 S 4 resulted in a decrease in the Cr(VI) reduction rates. It can be ascribed to the low photocatalytic activity of the pure CaIn 2 S 4 because of the slow separation of the photogenerated charge carriers. Moreover, the apparent quantum efficiencies (AQE) of the Cr(VI) photocatalytic reduction over the synthesized composites were also calculated, and the corresponding results are summarized in Table 1. As indicated in Table 1, the AQE of the 30% CaIn 2 S 4 /ZnIn 2 S 4 nanocomposite catalyst reached 6.6%, which presented higher than that of the pure ZnIn 2 S 4 (3.7%) or pure CaIn 2 S 4 (2.3%). In addition, the AQE for the Cr(VI) reduction increased gradually with the increase of the CaIn 2 S 4 constituent when the addition ratio of the CaIn 2 S 4 precursor was no more than 30%. However, further increasing the molar ratio of CaIn 2 S 4 led to a decrease in the AQE. This suggests that the heterostructured composites containing a suitable amount of CaIn 2 S 4 and ZnIn 2 S 4 contributed to improving optimally the photoactivity for the Cr(VI) reduction. Therefore, the optimal 30% CaIn 2 S 4 /ZnIn 2 S 4 has been chosen as a representative catalyst for the following studies.
To further clarify the influence of the heterostructure on the photocatalytic performance of the Cr(VI) reduction, the 30% CaIn 2 S 4 /ZnIn 2 S 4 sample was compared to its mechanical mixing counterpart sample 30% CaIn 2 S 4 + 70% ZnIn 2 S 4 . The photocatalytic activity of the Cr(VI) reduction over the physical mixture was much lower than that of the 30% CaIn 2 S 4 /ZnIn 2 S 4 composite obtained via the one-step hydrothermal method. This result shows that the heterojunction formed between the CaIn 2 S 4 and ZnIn 2 S 4 contributed to improving photocatalytic efficiency of the Cr(VI) reduction.
When using the optimal 30% CaIn 2 S 4 /ZnIn 2 S 4 as a photocatalyst, the change in the temporal absorption spectra of the DPC-Cr(VI) complex solution is illustrated in Figure 6B. The absorption peak at 540 nm belonging to the DPC-Cr(VI) complex decreased rapidly with the increase of light irradiation time, and it almost vanished after light illumination for 30 min. To gain more insight into the photocatalytic process, the total Cr ions concentrations over the 30% CaIn 2 S 4 /ZnIn 2 S 4 photocatalyst after the treated samples were measured by ICP emission spectrometer ( Figure S1, Supplementary Materials). As indicated in Figure S1, the initial concentration of Cr(VI) is 19.6 ppm. The measured total Cr ions concentration after 30 min light irradiation were found to be 19.1 ppm. This result demonstrates that almost no Cr(0) was produced in the present photocatalytic system. That is to say, Cr(VI) was primarily reduced to Cr(III) by the photogenerated electrons of the CaIn 2 S 4 /ZnIn 2 S 4 composites.
In addition, the photocatalytic experiments for the Cr(VI) reduction under different pH conditions with 30% CaIn 2 S 4 /ZnIn 2 S 4 were also carried out, and the corresponding results are presented in Figure 6C. In a photocatalytic system, the reduction efficiency of the aqueous Cr(VI) was greatly influenced by the pH value according to the previous reports [30]. Cr(VI) existed in two forms in alkaline and acid medium, respectively. CrO 4 2− is predominant in the alkaline medium, whereas Cr 2 O 7 2− plays a major role in the acid medium. The chemical redox reaction can be outlined as follows: 14H + + Cr 2 O 7 2− + 6e − → 2Cr 3+ + 7H 2 O (acid) (5) As depicted in Figure 6C, the reduction rate of Cr(VI) increased with the decrease of the pH value. About 38.7%, 58.8%, 80.6%, and 89.0% of Cr(VI) are reduced at pH 10, 8, 6, and 4 in the first 15 min, respectively. This could be due to the fact that the Cr(OH) 3 precipitate produced in the alkaline medium covers the activity sites of the photocatalysts [38]. The results demonstrate the acidic condition was more beneficial to the photocatalytic reduction of Cr(VI) over the synthesized CaIn 2 S 4 /ZnIn 2 S 4 composites.
Additionally, the controlled experiment by adding K 2 S 2 O 8 (the trapping agent of photo-generated electrons, 0.1 mmol) [39] into the photocatalytic system of the Cr(VI) reduction was performed, and the corresponding result is depicted in Figure 6D. It can be seen clearly that the Cr(VI) reduction over the optimal photocatalyst 30% CaIn 2 S 4 /ZnIn 2 S 4 hardly occurs in the presence of K 2 S 2 O 8 , which indicates that the reduction of Cr(VI) is conducted by photogenerated electrons under simulated sunlight irradiation. These results also indicate that a suitable amount of CaIn 2 S 4 can effectively improve the photocatalytic performance of ZnIn 2 S 4 toward the Cr(VI) reduction.

Catalytic Stability
The stability of a given photocatalyst was also an important factor in the practical application [40]. Furthermore, narrow bandgap semiconductors are generally more unstable when exposed to the sunlight illumination [24]. To study the effect of the CaIn 2 S 4 /ZnIn 2 S 4 heterostructure on the catalytic stability, the cycle experiments of the Cr(VI) reduction were carried out and compared by using ZnIn 2 S 4 and 30% CaIn 2 S 4 /ZnIn 2 S 4 as photocatalysts under simulated sunlight irradiation (Figure 7). The results indicate that for the pristine ZnIn 2 S 4 , the reduction efficiency of Cr(VI) decreased about 20.9% after five repeated uses, demonstrating that ZnIn 2 S 4 was unstable to some extent under simulated sunlight illumination. In comparison, a relatively low decrease in the reduction efficiency of Cr(VI) over 30% CaIn 2 S 4 /ZnIn 2 S 4 was observed, and only a ca. 2.6% decrease of the reduction efficiency is obtained after reusing five cycles. This suggests that the sunlight stability of ZnIn 2 S 4 can be improved through forming the heterojunction composites with CaIn 2 S 4 . Additionally, the XPS analysis of the used 30% CaIn 2 S 4 /ZnIn 2 S 4 catalyst was also performed, and the obtained results are displayed in Figure S2 of the Supplementary Materials. As can be found from Figure S2, the surface element composition and the chemical state of the 30% CaIn 2 S 4 /ZnIn 2 S 4 sample before and after the photocatalytic reaction show no obvious difference. The results confirm the superior stability of the CaIn 2 S 4 /ZnIn 2 S 4 composite photocatalysts under simulated sunlight, which is more promising for the practical photocatalytic applications in environmental restoration. stability of ZnIn2S4 can be improved through forming the heterojunction composites with CaIn2S4. Additionally, the XPS analysis of the used 30% CaIn2S4/ZnIn2S4 catalyst was also performed, and the obtained results are displayed in Figure S2 of the Supplementary Materials. As can be found from Figure S2, the surface element composition and the chemical state of the 30% CaIn2S4/ZnIn2S4 sample before and after the photocatalytic reaction show no obvious difference. The results confirm the superior stability of the CaIn2S4/ZnIn2S4 composite photocatalysts under simulated sunlight, which is more promising for the practical photocatalytic applications in environmental restoration.

Enhancement Mechanism of Photocatalytic Activity and Stability
On the basis of the above discussion and results, it is obvious that coupling a suitable amount of CaIn2S4 can dramatically improve the photocatalytic performance for the Cr(VI) reduction, including the reduction efficiency and cycling stability of the CaIn2S4/ZnIn2S4 composites. The remarkably enhanced photocatalytic activity of the CaIn2S4/ZnIn2S4 samples, compared with the pure CaIn2S4 and ZnIn2S4, can be ascribed to the effective separation of the photogenerated electron/hole pairs due to the forming of flower-like heterostructures between ZnIn2S4 and CaIn2S4. Generally, the photoluminescence spectrum (PLs) is considered as a vital technology to study the migration and fate of photoinduced charge carriers [39,41]. Higher intensities of PL signals usually represent higher recombination rates of photogenerated charge carriers, thus resulting in a lower photocatalytic performance. The comparison of the PL spectrum for the pristine ZnIn2S4 and 30% CaIn2S4/ZnIn2S4 with an excitation wavelength of 330 nm is presented in Figure 8. The pristine ZnIn2S4 showed emissions at 520, 571, and 656 nm. The strong emission peak centered at 520 nm for the pristine ZnIn2S4 was attributed to the intrinsic luminescence of ZnIn2S4. The relatively weak PL peaks at 571, and 656 nm can be ascribed to the surface state emissions, which were mainly caused by the surface defects in the ZnIn2S4 structure. In comparison with the pristine ZnIn2S4, there was no new emission signal in the PL spectrum of 30% CaIn2S4/ZnIn2S4, but the intensities of the PL peaks decreased obviously. The results imply that the photogenerated electrons and holes can transfer effectively between CaIn2S4 and ZnIn2S4, thus suppressing the recombination of charge carriers. This can be a primary reason for the CaIn2S4/ZnIn2S4 composites possessing excellent photocatalytic reduction performance under simulated sunlight irradiation. Photocatalytic stability tests of ZnIn 2 S 4 and 30% CaIn 2 S 4 /ZnIn 2 S 4 toward the Cr(VI) reduction.

Enhancement Mechanism of Photocatalytic Activity and Stability
On the basis of the above discussion and results, it is obvious that coupling a suitable amount of CaIn 2 S 4 can dramatically improve the photocatalytic performance for the Cr(VI) reduction, including the reduction efficiency and cycling stability of the CaIn 2 S 4 /ZnIn 2 S 4 composites. The remarkably enhanced photocatalytic activity of the CaIn 2 S 4 /ZnIn 2 S 4 samples, compared with the pure CaIn 2 S 4 and ZnIn 2 S 4 , can be ascribed to the effective separation of the photogenerated electron/hole pairs due to the forming of flower-like heterostructures between ZnIn 2 S 4 and CaIn 2 S 4 . Generally, the photoluminescence spectrum (PLs) is considered as a vital technology to study the migration and fate of photoinduced charge carriers [39,41]. Higher intensities of PL signals usually represent higher recombination rates of photogenerated charge carriers, thus resulting in a lower photocatalytic performance. The comparison of the PL spectrum for the pristine ZnIn 2 S 4 and 30% CaIn 2 S 4 /ZnIn 2 S 4 with an excitation wavelength of 330 nm is presented in Figure 8. The pristine ZnIn 2 S 4 showed emissions at 520, 571, and 656 nm. The strong emission peak centered at 520 nm for the pristine ZnIn 2 S 4 was attributed to the intrinsic luminescence of ZnIn 2 S 4 . The relatively weak PL peaks at 571, and 656 nm can be ascribed to the surface state emissions, which were mainly caused by the surface defects in the ZnIn 2 S 4 structure. In comparison with the pristine ZnIn 2 S 4 , there was no new emission signal in the PL spectrum of 30% CaIn 2 S 4 /ZnIn 2 S 4 , but the intensities of the PL peaks decreased obviously. The results imply that the photogenerated electrons and holes can transfer effectively between CaIn 2 S 4 and ZnIn 2 S 4 , thus suppressing the recombination of charge carriers. This can be a primary reason for the CaIn 2 S 4 /ZnIn 2 S 4 composites possessing excellent photocatalytic reduction performance under simulated sunlight irradiation. Most of the heterostructured nanocomposites follow the bidirectional charge transfer mechanism [42,43]. For instance, Kumar and his colleagues have reported on the photoinduced electrons enriched on the conduction band of Ag3PO4 and holes on the valence band of g-C3N4, which was conducted through a bidirectional charge transfer process between Ag3PO4 and g-C3N4 [42]. As for the as-prepared CaIn2S4/ZnIn2S4 composites, the aforementioned bidirectional charge transfer was also the primary charge migration process. Based on the bandgap energies of CaIn2S4 and ZnIn2S4 estimated from Figure 5 and Equations (2) and (3), the energy band structure diagram of CaIn2S4 and ZnIn2S4 can be schematically illustrated, as shown in Figure 9. Under light illumination, the components of CaIn2S4 and ZnIn2S4 were simultaneously excited, generating photoinduced electron/hole pairs. Due to the more negative conduction-band edge of CaIn2S4 (−1.12 eV) than that of ZnIn2S4 (−0.85 eV), the photogenerated electrons prefer to transfer from the CB of CaIn2S4 to ZnIn2S4, whereas the photogenerated holes on the more positive VB of ZnIn2S4 (+1.58 eV) would migrate to that of CaIn2S4 (+0.91 eV). This bidirectional charge transfer process results in efficient separation of photogenerated electron/hole pairs in the synthesized CaIn2S4/ZnIn2S4 composites. To further investigate the important function of heterostructures in enhancing the separation of charge carriers, the transient photocurrent responses were measured over working electrodes made of the pure CaIn2S4, ZnIn2S4, and 30% CaIn2S4/ZnIn2S4 composite. As depicted in Figure 10, the fast and steady photocurrent response can be detected for each light-on and light-off cycle over the pure CaIn2S4, ZnIn2S4, and 30% CaIn2S4/ZnIn2S4 composite. The pristine CaIn2S4 exhibited a very low photocurrent density, whereas the pure ZnIn2S4 showed a relatively higher photocurrent than that of CaIn2S4 under simulated sunlight irradiation. This could be due to the fact that CaIn2S4 Most of the heterostructured nanocomposites follow the bidirectional charge transfer mechanism [42,43]. For instance, Kumar and his colleagues have reported on the photoinduced electrons enriched on the conduction band of Ag 3 PO 4 and holes on the valence band of g-C 3 N 4 , which was conducted through a bidirectional charge transfer process between Ag 3 PO 4 and g-C 3 N 4 [42]. As for the as-prepared CaIn 2 S 4 /ZnIn 2 S 4 composites, the aforementioned bidirectional charge transfer was also the primary charge migration process. Based on the bandgap energies of CaIn 2 S 4 and ZnIn 2 S 4 estimated from Figure 5 and Equations (2) and (3), the energy band structure diagram of CaIn 2 S 4 and ZnIn 2 S 4 can be schematically illustrated, as shown in Figure 9. Under light illumination, the components of CaIn 2 S 4 and ZnIn 2 S 4 were simultaneously excited, generating photoinduced electron/hole pairs. Due to the more negative conduction-band edge of CaIn 2 S 4 (−1.12 eV) than that of ZnIn 2 S 4 (−0.85 eV), the photogenerated electrons prefer to transfer from the CB of CaIn 2 S 4 to ZnIn 2 S 4 , whereas the photogenerated holes on the more positive VB of ZnIn 2 S 4 (+1.58 eV) would migrate to that of CaIn 2 S 4 (+0.91 eV). This bidirectional charge transfer process results in efficient separation of photogenerated electron/hole pairs in the synthesized CaIn 2 S 4 /ZnIn 2 S 4 composites. Most of the heterostructured nanocomposites follow the bidirectional charge transfer mechanism [42,43]. For instance, Kumar and his colleagues have reported on the photoinduced electrons enriched on the conduction band of Ag3PO4 and holes on the valence band of g-C3N4, which was conducted through a bidirectional charge transfer process between Ag3PO4 and g-C3N4 [42]. As for the as-prepared CaIn2S4/ZnIn2S4 composites, the aforementioned bidirectional charge transfer was also the primary charge migration process. Based on the bandgap energies of CaIn2S4 and ZnIn2S4 estimated from Figure 5 and Equations (2) and (3), the energy band structure diagram of CaIn2S4 and ZnIn2S4 can be schematically illustrated, as shown in Figure 9. Under light illumination, the components of CaIn2S4 and ZnIn2S4 were simultaneously excited, generating photoinduced electron/hole pairs. Due to the more negative conduction-band edge of CaIn2S4 (−1.12 eV) than that of ZnIn2S4 (−0.85 eV), the photogenerated electrons prefer to transfer from the CB of CaIn2S4 to ZnIn2S4, whereas the photogenerated holes on the more positive VB of ZnIn2S4 (+1.58 eV) would migrate to that of CaIn2S4 (+0.91 eV). This bidirectional charge transfer process results in efficient separation of photogenerated electron/hole pairs in the synthesized CaIn2S4/ZnIn2S4 composites. To further investigate the important function of heterostructures in enhancing the separation of charge carriers, the transient photocurrent responses were measured over working electrodes made of the pure CaIn2S4, ZnIn2S4, and 30% CaIn2S4/ZnIn2S4 composite. As depicted in Figure 10, the fast and steady photocurrent response can be detected for each light-on and light-off cycle over the pure CaIn2S4, ZnIn2S4, and 30% CaIn2S4/ZnIn2S4 composite. The pristine CaIn2S4 exhibited a very low photocurrent density, whereas the pure ZnIn2S4 showed a relatively higher photocurrent than that of CaIn2S4 under simulated sunlight irradiation. This could be due to the fact that CaIn2S4 To further investigate the important function of heterostructures in enhancing the separation of charge carriers, the transient photocurrent responses were measured over working electrodes made of the pure CaIn 2 S 4 , ZnIn 2 S 4 , and 30% CaIn 2 S 4 /ZnIn 2 S 4 composite. As depicted in Figure 10, the fast and steady photocurrent response can be detected for each light-on and light-off cycle over the pure CaIn 2 S 4 , ZnIn 2 S 4 , and 30% CaIn 2 S 4 /ZnIn 2 S 4 composite. The pristine CaIn 2 S 4 exhibited a very low photocurrent density, whereas the pure ZnIn 2 S 4 showed a relatively higher photocurrent than that of CaIn 2 S 4 under simulated sunlight irradiation. This could be due to the fact that CaIn 2 S 4 possesses a narrower bandgap than ZnIn 2 S 4 , which is unfavorable for the effective separation of photoinduced electron/hole pairs. This would result in the short survival time of photogenerated electrons and weak photocurrent density. However, the composite sample of 30% CaIn 2 S 4 /ZnIn 2 S 4 exhibits a dramatically enhanced photocurrent density compared with that of the pure ZnIn 2 S 4 and CaIn 2 S 4 , which further substantiates the efficient separation of the photoinduced electron/hole pairs in the obtained CaIn 2 S 4 /ZnIn 2 S 4 composites. possesses a narrower bandgap than ZnIn2S4, which is unfavorable for the effective separation of photoinduced electron/hole pairs. This would result in the short survival time of photogenerated electrons and weak photocurrent density. However, the composite sample of 30% CaIn2S4/ZnIn2S4 exhibits a dramatically enhanced photocurrent density compared with that of the pure ZnIn2S4 and CaIn2S4, which further substantiates the efficient separation of the photoinduced electron/hole pairs in the obtained CaIn2S4/ZnIn2S4 composites. Based on the above experimental results and the charge transfer process depicted in Figure 9, a possible enhancement mechanism of photocatalytic activity and stability for the Cr(VI) reduction can be proposed. Under simulated sunlight irradiation, the components of CaIn2S4 and ZnIn2S4 are excited to produce photoinduced holes and electrons. Owing to the relatively high recombination rate of photogenerated electron/hole pairs, the pristine CaIn2S4 and ZnIn2S4 exhibited low photocatalytic performance. Due to the well-matched energy band structures and the intimate interfacial contact between CaIn2S4 and ZnIn2S4 in the as-synthesized composites, the photoinduced electrons located on the CB of CaIn2S4 can easily migrate to that of ZnIn2S4, and on the contrary, the photoinduced holes on the VB of ZnIn2S4 spontaneously transfer to that of CaIn2S4 ( Figure 9). These photoelectrons (e − ) accumulated on the CB of ZnIn2S4 possess a strong reduction ability (−0.85 eV vs. NHE (normal hydrogen electrode)), which can reduce toxic Cr(VI) to Cr(III) (ECr(VI)/Cr(III) = +0.55 eV vs. NHE) [44,45]. The photoholes located on the VB of CaIn2S4 react with the sacrificial reagents immediately. The heterostructured composites formed between CaIn2S4 and ZnIn2S4 effectively prevent the recombination of photogenerated electrons and holes, and thus, the photocatalytic activities of the Cr(VI) reduction are enhanced greatly. Meanwhile, the effective transfer of photogenerated holes from the VB of ZnIn2S4 to that of CaIn2S4 is beneficial for preventing the oxidation of S 2− by holes, which significantly improves the photostability of ZnIn2S4 in the composite catalysts.

Conclusions
In short, the CaIn2S4/ZnIn2S4 composite photocatalysts were successfully prepared through a one-step hydrothermal process. XRD patterns show that the as-synthesized flower-like composites consist of hexagonal phase ZnIn2S4 and cubic phase CaIn2S4. Compared with the pristine ZnIn2S4, the heterostructured composites CaIn2S4/ZnIn2S4 show significantly improved photocatalytic activity and stability for the Cr(VI) reduction under simulated sunlight illumination. The molar content of CaIn2S4 has a great influence on the photocatalytic activity of the CaIn2S4/ZnIn2S4 composites, and 30% CaIn2S4/ZnIn2S4 exhibits the optimal photocatalytic performance for the Cr(VI) reduction. A possible mechanism of the photogenerated charge transfer was proposed to illustrate the superior photocatalytic performance and photostability of the CaIn2S4/ZnIn2S4 composite  Figure 10.
Photocurrent spectra of the as-synthesized pure CaIn 2 S 4 , ZnIn 2 S 4 , and 30% CaIn 2 S 4 /ZnIn 2 S 4 samples under simulated sunlight irradiation with 20 s light on/off cycles. Based on the above experimental results and the charge transfer process depicted in Figure 9, a possible enhancement mechanism of photocatalytic activity and stability for the Cr(VI) reduction can be proposed. Under simulated sunlight irradiation, the components of CaIn 2 S 4 and ZnIn 2 S 4 are excited to produce photoinduced holes and electrons. Owing to the relatively high recombination rate of photogenerated electron/hole pairs, the pristine CaIn 2 S 4 and ZnIn 2 S 4 exhibited low photocatalytic performance. Due to the well-matched energy band structures and the intimate interfacial contact between CaIn 2 S 4 and ZnIn 2 S 4 in the as-synthesized composites, the photoinduced electrons located on the CB of CaIn 2 S 4 can easily migrate to that of ZnIn 2 S 4 , and on the contrary, the photoinduced holes on the VB of ZnIn 2 S 4 spontaneously transfer to that of CaIn 2 S 4 ( Figure 9). These photoelectrons (e − ) accumulated on the CB of ZnIn 2 S 4 possess a strong reduction ability (−0.85 eV vs. NHE (normal hydrogen electrode)), which can reduce toxic Cr(VI) to Cr(III) (E Cr(VI)/Cr(III) = +0.55 eV vs. NHE) [44,45]. The photoholes located on the VB of CaIn 2 S 4 react with the sacrificial reagents immediately. The heterostructured composites formed between CaIn 2 S 4 and ZnIn 2 S 4 effectively prevent the recombination of photogenerated electrons and holes, and thus, the photocatalytic activities of the Cr(VI) reduction are enhanced greatly. Meanwhile, the effective transfer of photogenerated holes from the VB of ZnIn 2 S 4 to that of CaIn 2 S 4 is beneficial for preventing the oxidation of S 2− by holes, which significantly improves the photostability of ZnIn 2 S 4 in the composite catalysts.

Conclusions
In short, the CaIn 2 S 4 /ZnIn 2 S 4 composite photocatalysts were successfully prepared through a one-step hydrothermal process. XRD patterns show that the as-synthesized flower-like composites consist of hexagonal phase ZnIn 2 S 4 and cubic phase CaIn 2 S 4 . Compared with the pristine ZnIn 2 S 4 , the heterostructured composites CaIn 2 S 4 /ZnIn 2 S 4 show significantly improved photocatalytic activity and stability for the Cr(VI) reduction under simulated sunlight illumination. The molar content of CaIn 2 S 4 has a great influence on the photocatalytic activity of the CaIn 2 S 4 /ZnIn 2 S 4 composites, and 30% CaIn 2 S 4 /ZnIn 2 S 4 exhibits the optimal photocatalytic performance for the Cr(VI) reduction.
A possible mechanism of the photogenerated charge transfer was proposed to illustrate the superior photocatalytic performance and photostability of the CaIn 2 S 4 /ZnIn 2 S 4 composite catalysts. This study is of great importance in the design and synthesis of heterostructured sulfide composites with excellent photocatalytic performance and consistent stability toward the elimination of toxic metal ions in water.