Enhanced Photocatalytic H2 Evolution over ZnIn2S4 Flower-Like Microspheres Doped with Black Phosphorus Quantum Dots

In this work, black phosphorus quantum dots (BPQDs) were decorated on hexagonal ZnIn2S4 flower-like microspheres to form zero-dimensional/two-dimensional (0D/2D) structures. Interface interactions between the BPQDs and ZnIn2S4 resulted in optimum effective charge transfer, thereby improving the photocatalytic performance of the material. Thus, the 0.2% BPQD–ZnIn2S4 sample showed 30% higher H2 evolution rates compared to pure ZnIn2S4. This study provides a simple route for the synthesis of photocatalysts. The results obtained herein can pave the way for designing effective catalysts for solar-to-chemical energy conversion and feasible approaches to obtain cheap, clean, and efficient photocatalysts.


Introduction
H 2 is regarded as the most promising energy vector in the 21st century. Photocatalytic H 2 evolution from water via solar irradiation is an attractive and sustainable solution for addressing the energy supply issue [1][2][3]. Since the photocatalytic splitting of water into H 2 over TiO 2 under solar irradiation was first discovered by Fujishima in 1972 [4], this reaction received broad attention. Thus, numerous efforts were made to develop non-toxic, clean, and efficient semiconductor photocatalysts. In this sense, a series of photocatalysts with enhanced photocatalytic H 2 evolution activity such as CdS [5][6][7] and graphitic carbon nitride (g-C 3 N 4 ) [8][9][10][11], among others, were developed so far.
Metal sulfides attract much attention owing to their suitable band position and strong visible-light absorption characteristics [12]. Among the ternary metal sulfides, chalcogenide ZnIn 2 S 4 combines a suitable band gap, high chemical stability, and excellent photocatalytic activity, making it a suitable photocatalyst [13][14][15]. However, ZnIn 2 S 4 suffers from a quick recombination of photogenerated electron-hole pairs and low migration ability for the photoexcited charge carriers. These are, in fact, the fundamental drawbacks of most semiconductors [16]. In order to optimize the photocatalytic performance of these materials, much effort was made for studying preparation methods leading to suitable particle size, structure, and morphology [17][18][19]. In addition, the photocatalytic activity depends significantly on the migration of the electron-hole pairs from the conduction band (CB) to the valence band (VB). Thus, the photocatalytic performance of ZnIn 2 S 4 can be mainly improved by expanding its light absorption region and by eliminating the recombination centers of photoinduced

Preparation
All chemicals were of analytical grade and used as received without further purification.

Preparation of Hexagonal ZnIn 2 S 4 Flower-Like Microspheres
In a typical procedure, 1.0 mmol of ZnCl 2 , 2.0 mmol of InCl 3 , and 6.0 mmol of thioacetamide (TAA) were dissolved in 80 mL of deionized water. The mixed solution was vigorously stirred for 30 min and then transferred to a 150-mL Teflon-lined autoclave. The autoclave was sealed, kept at 80 • C for 12 h, and then cooled to room temperature naturally. A yellow precipitate was filtered and washed with deionized water and absolute ethanol several times. Then, the solid was oven-dried at 60 • C overnight. The sample was denoted as ZIS.

Preparation of a Dispersion of BPQDs
A dispersion of BPQDs was synthesized via solvent exfoliation with dimethylformamide (DMF). Firstly, 10 mg of bulk BP was added to 10 mL of DMF. The dispersion was sonicated for 6 h on an ultrasonic homogenizer (Scientz, JY92-II N, Ningbo, China) under ice cooling. After that, the dispersion was centrifuged at 5000 rpm for 20 min (to remove the non-exfoliated bulk BP) followed by centrifugation at 13,000 rpm for another 30 min to remove the BP flakes. The resulting material was a BPQD/DMF dispersion.

Preparation of BPQD-ZIS Materials
BPQD-ZIS materials were prepared by adding 200 mg of ZIS to DMF, followed by ultrasonication of the mixture to form a homogeneous dispersion. Subsequently, the as-prepared BPQD/DMF dispersion was added to a ZIS dispersion and sonicated for 2 h, and the mixture was stirred overnight. The as-obtained powders were collected by high-speed centrifugation, washed with deionized water and ethanol thoroughly, and finally vacuum-dried at 40 • C overnight to obtain BPQD-ZIS. Several Nanomaterials 2019, 9, 1266 3 of 13 BPQD-ZIS materials with various weight ratios were prepared by adding different amounts of the BPQD dispersion to the ZIS dispersion.

Photoelectrochemical Measurements
Photoelectrochemical measurements of the obtained products were carried out on an electrochemical analyzer (CHI660E) in a standard three-electrode system using the prepared samples as the working electrodes with an active area of 1 cm 2 , a Pt wire as the counter electrode, and saturated Ag/AgCl as the reference electrode. An aqueous Na 2 SO 4 (0.2 M) solution was used as the electrolyte. The electrochemical impedance spectra (EIS) were obtained by applying an alternate current amplitude of 5.0 mV in a frequency varying from 100 kHz to 0.01 Hz.

Photocatalytic H 2 Evolution
Typically, 50 mg of catalyst powder was added to a cylinder reactor containing an aqueous solution (100 mL) containing 0.35 M Na 2 S and 0.25 M Na 2 SO 3 and sealed with a rubber septum. Before the reaction, Ar purges were carried out through the solution to remove the O 2 and N 2 dissolved in the water. The reaction was carried out by irradiating the suspension with simulated sunlight with a 300-W Xe lamp (Perfect Light, PLS-SXE300, Beijing, China) equipped with an optical filter AM 1.5 (Perfect Light, Beijing, China). The gas produced over the catalyst was collected and analyzed by gas chromatography (GC, Shiweipuxin GC-7806, Shiweipx, Beijing, China).

Characterization of the Photocatalysts
The XRD patterns of the pure ZnIn 2 S 4 and hybrid BPQD catalysts (BPQD loadings of 0.1, 0.  [39], respectively. These results indicated that the presence of BPQDs did not alter the crystal structure of ZnIn 2 S 4 . The low loadings of BPQDs prevented us from observing apparent peaks corresponding to BPQDs in the BPQD-ZIS samples. The morphology of the prepared BPQD-ZIS was confirmed by SEM. As shown in Figure 2, pure ZnIn 2 S 4 contained flower-like microspheres of ca. 2-5 µm in diameter. As revealed by TEM (Figure 2d), the flower-like microspheres of the synthesized BPQD-ZIS materials contained numerous ultrathin nanosheets. The presence of these nanosheets rendered ZnIn 2 S 4 with a unique layered structure. Thus, the ZnIn 2 S 4 flower-like microspheres increased the contact area for BPQD decoration. BPQDs were observed at the edges of the flower-like microspheres, revealing that BPQD-doped ZnIn 2 S 4 hybrids were successfully synthesized. A lattice spacing of 0.32 nm was determined ( Figure 2f) and assigned to the (102) facet of hexagonal ZnIn 2 S 4 [15]. The inter-planar spacing of ca. 0.33 nm corresponded to the (021) facet of BPQDs [40] in the hybrid material. These results demonstrated that BPQDs were introduced into the ZnIn 2 S 4 phase, generating a good contact between the BPQDs and ZnIn 2 S 4 in the BPQD-ZIS hybrids. Elemental mapping of the BPQD-ZIS samples revealed the presence of P, Zn, In, and S, confirming that the BPQD-ZIS hybrids were successfully prepared. corresponded to the (021) facet of BPQDs [40] in the hybrid material. These results demonstrated that BPQDs were introduced into the ZnIn2S4 phase, generating a good contact between the BPQDs and ZnIn2S4 in the BPQD-ZIS hybrids. Elemental mapping of the BPQD-ZIS samples revealed the presence of P, Zn, In, and S, confirming that the BPQD-ZIS hybrids were successfully prepared.  To gain insight into the interfacial interaction between ZnIn2S4 and BPQDs, XPS measurements (Zn 2p, In 3d, S 2p, and P 2p) were carried out. The full survey scanned spectra (Figure 3a  corresponded to the (021) facet of BPQDs [40] in the hybrid material. These results demonstrated that BPQDs were introduced into the ZnIn2S4 phase, generating a good contact between the BPQDs and ZnIn2S4 in the BPQD-ZIS hybrids. Elemental mapping of the BPQD-ZIS samples revealed the presence of P, Zn, In, and S, confirming that the BPQD-ZIS hybrids were successfully prepared.  To gain insight into the interfacial interaction between ZnIn2S4 and BPQDs, XPS measurements (Zn 2p, In 3d, S 2p, and P 2p) were carried out. The full survey scanned spectra (Figure 3a  To gain insight into the interfacial interaction between ZnIn 2 S 4 and BPQDs, XPS measurements (Zn 2p, In 3d, S 2p, and P 2p) were carried out. The full survey scanned spectra (Figure 3a) of ZnIn 2 S 4 and the BPQD-ZIS hybrids contained the Zn 2p, In 3d, and S 2p bands. The high-resolution XPS Zn 2p band of ZnIn 2 S 4 ( Figure 3b) contained two peaks at 1044.88 and 1022.34 eV, corresponding to the Zn 2p 1/2 and Zn 2p 3/2 transitions of Zn 2+ species, respectively. The peaks at 452.61 and 445.1 eV (Figure 3c) corresponded to In 3d 3/2 and In 3d 5/2 transitions of In 3+ species, respectively. The peaks at 163.18 and 161.82 eV (Figure 3d) corresponded to the S 2p band of ZnIn 2 S 4 , and were attributed to S 2p 1/2 and S 2p 3/2 transitions [39], respectively. However, the Zn 2p, In 3d, and S 2p bands shifted to higher binding energies for the BPQD-ZIS hybrid material compared to ZnIn 2 S 4 . The electronic coupling between ZnIn 2 S 4 and BPQDs upon introduction of BPQDs can account for these shifts. Furthermore, as shown in Figure 3e, the three characteristic P 2p peaks at ca. 129.8, 130.6, and 133.6 eV were observed for BPQDs [37] after curve-fitting, and these peaks corresponded to P 2p 3/2 , P 2p 1/2 , and oxidized phosphorus (POx), respectively. BP crystals are very stable in air. However, when BP is in the form of QDs, it becomes especially sensitive to water and oxygen. Since our experiments were carried out in air, BPQDs were unavoidably oxidized during the process. The high-resolution P 2p XPS spectra of the BPQD-ZIS samples further confirmed the formation of the BPQD-ZIS hybrid system. Zn 2p1/2 and Zn 2p3/2 transitions of Zn 2+ species, respectively. The peaks at 452.61 and 445.1 eV ( Figure  3c) corresponded to In 3d3/2 and In 3d5/2 transitions of In 3+ species, respectively. The peaks at 163.18 and 161.82 eV (Figure 3d) corresponded to the S 2p band of ZnIn2S4, and were attributed to S 2p1/2 and S 2p3/2 transitions [39], respectively. However, the Zn 2p, In 3d, and S 2p bands shifted to higher binding energies for the BPQD-ZIS hybrid material compared to ZnIn2S4. The electronic coupling between ZnIn2S4 and BPQDs upon introduction of BPQDs can account for these shifts. Furthermore, as shown in Figure 3e, the three characteristic P 2p peaks at ca. 129.8, 130.6, and 133.6 eV were observed for BPQDs [37] after curve-fitting, and these peaks corresponded to P 2p3/2, P 2p1/2, and oxidized phosphorus (POx), respectively. BP crystals are very stable in air. However, when BP is in the form of QDs, it becomes especially sensitive to water and oxygen. Since our experiments were carried out in air, BPQDs were unavoidably oxidized during the process. The high-resolution P 2p XPS spectra of the BPQD-ZIS samples further confirmed the formation of the BPQD-ZIS hybrid system. The Brunauer-Emmett-Teller (BET) specific surface area of pure ZIS and 0.2% BPQD-ZIS samples were investigated by N2 adsorption-desorption measurements. As can be clearly seen in Figure 4, the curved shape of the samples belonged to a type IV isotherm with hysteresis loops characteristic of a mesoporous structure. The BET specific surface areas of pure ZIS and 0.2% BPQD-ZIS were calculated to be 116. 35 and 118.52 m 2 •g −1 , suggesting that the incorporation of BPQDs had negligible influence on the surface area of ZIS. The slight larger surface area of the 0.2% BPQD-ZIS sample may be due to testing errors.  The Brunauer-Emmett-Teller (BET) specific surface area of pure ZIS and 0.2% BPQD-ZIS samples were investigated by N 2 adsorption-desorption measurements. As can be clearly seen in Figure 4, the curved shape of the samples belonged to a type IV isotherm with hysteresis loops characteristic of a mesoporous structure. The BET specific surface areas of pure ZIS and 0.2% BPQD-ZIS were calculated to be 116. 35 and 118.52 m 2 ·g −1 , suggesting that the incorporation of BPQDs had negligible influence on the surface area of ZIS. The slight larger surface area of the 0.2% BPQD-ZIS sample may be due to testing errors. FT-IR spectroscopy was conducted to further characterize the prepared samples and the results are shown in Figure 5. Pure ZIS and the BPQD-ZIS samples showed similar FT-IR bands. The peaks at about 1620 and 1385 cm −1 corresponded to the hydroxyl group and surface-absorbed water molecules, and the absorption peak found at 1100 cm −1 was assigned to the absorbed CO2. The characteristic peaks of BPQDs were not observed [13,41], while the peak located at around 640 cm −1 could be assigned to the stretching vibrations of In-S [42], which gradually enhanced with increased BPQDs. The reason for this phenomenon is that the incorporation of BPQDs may influence the stretching vibrations of In-S, resulting in increased peak intensity. These results further confirmed the contact between the BPQDs and ZnIn2S4 and the formation of BPQD-ZIS hybrids.  FT-IR spectroscopy was conducted to further characterize the prepared samples and the results are shown in Figure 5. Pure ZIS and the BPQD-ZIS samples showed similar FT-IR bands. The peaks at about 1620 and 1385 cm −1 corresponded to the hydroxyl group and surface-absorbed water molecules, and the absorption peak found at 1100 cm −1 was assigned to the absorbed CO 2 . The characteristic peaks of BPQDs were not observed [13,41], while the peak located at around 640 cm −1 could be assigned to the stretching vibrations of In-S [42], which gradually enhanced with increased BPQDs. The reason for this phenomenon is that the incorporation of BPQDs may influence the stretching vibrations of In-S, resulting in increased peak intensity. These results further confirmed the contact between the BPQDs and ZnIn 2 S 4 and the formation of BPQD-ZIS hybrids. FT-IR spectroscopy was conducted to further characterize the prepared samples and the results are shown in Figure 5. Pure ZIS and the BPQD-ZIS samples showed similar FT-IR bands. The peaks at about 1620 and 1385 cm −1 corresponded to the hydroxyl group and surface-absorbed water molecules, and the absorption peak found at 1100 cm −1 was assigned to the absorbed CO2. The characteristic peaks of BPQDs were not observed [13,41], while the peak located at around 640 cm −1 could be assigned to the stretching vibrations of In-S [42], which gradually enhanced with increased BPQDs. The reason for this phenomenon is that the incorporation of BPQDs may influence the stretching vibrations of In-S, resulting in increased peak intensity. These results further confirmed the contact between the BPQDs and ZnIn2S4 and the formation of BPQD-ZIS hybrids. DRS was used to gain insight into the light absorption properties of ZnIn2S4 and the hybrid BPQD-ZIS materials. As shown in Figure 6a, all samples showed nearly similar absorption spectra. The spectra of the hybrid BPQD-ZIS materials red-shifted, and this can be ascribed to absorption of light by the BPQDs.  The spectra of the hybrid BPQD-ZIS materials red-shifted, and this can be ascribed to absorption of light by the BPQDs. In order to determine the band gap (Eg) of the samples, the Kubelka-Munk function was used.
The results were obtained by plotting (E × F(R)) 1/2 with respect to E, where R is the reflectivity, and A is the absorbance. As displayed in Figure 6b, these samples had a similar band gap. After the incorporation of BPQDs, the band gap of the samples slightly narrowed, whereby the results of pure ZIS, 0.1% BPQD-ZIS, 0.2% BPQD-ZIS, 0.3% BPQD-ZIS, and 0.4% BPQD-ZIS were 2.29, 2.26, 2.25, 2.28, and 2.26 eV, respectively. In addition, the band gap of BPQDs was also measured to be 2.68 eV, as shown in Figure 6c,d.

Photocatalytic H2 Evolution Performance of the Photocatalysts
The photocatalytic H2 production activities of ZnIn2S4 doped with different loadings of BPQDs were determined under simulated sunlight irradiation. As shown in Figure 7a, pure ZnIn2S4 released H2 (ca. 2786.45 μmol•g −1 •h −1 ) in aqueous solution (Na2S 0.35 M/Na2SO3 0.25 M). This may be ascribed to the fast recombination of the photoinduced electron-hole pairs between the CB and VB of ZnIn2S4. Moreover, 0.2 wt.% was found to be the optimal BPQD loading, resulting in photocatalytic H2 rates 1.3 times higher (3675.9 μmol•g −1 •h −1 ) compared to pure ZnIn2S4. The total yield of H2 increased significantly for the hybrid samples compared to pure ZnIn2S4. These results confirmed that decoration of BPQDs on ZnIn2S4 improved the photocatalytic activity of the materials. A further increase in the BPQD loading above 0.2 wt.% resulted in lower H2 production rates, probably because a fraction of the BPQDs acted as photo-recombination centers.
To confirm the positive effect of BPQDs in enhancing the photocatalytic H2 production activity of ZIS, the photocatalytic H2 production values of 0.2% BPQD-ZIS in this work were compared with the related literature, listed in Table 1. It can be clearly found that 0.2% BPQD-ZIS in this work In order to determine the band gap (Eg) of the samples, the Kubelka-Munk function was used.
The results were obtained by plotting (E × F(R)) 1/2 with respect to E, where R is the reflectivity, and A is the absorbance. As displayed in Figure 6b, these samples had a similar band gap. After the incorporation of BPQDs, the band gap of the samples slightly narrowed, whereby the results of pure ZIS, 0.1% BPQD-ZIS, 0.2% BPQD-ZIS, 0.3% BPQD-ZIS, and 0.4% BPQD-ZIS were 2.29, 2.26, 2.25, 2.28, and 2.26 eV, respectively. In addition, the band gap of BPQDs was also measured to be 2.68 eV, as shown in Figure 6c,d.

Photocatalytic H 2 Evolution Performance of the Photocatalysts
The photocatalytic H 2 production activities of ZnIn 2 S 4 doped with different loadings of BPQDs were determined under simulated sunlight irradiation. As shown in Figure 7a, pure ZnIn 2 S 4 released H 2 (ca. 2786.45 µmol·g −1 ·h −1 ) in aqueous solution (Na 2 S 0.35 M/Na 2 SO 3 0.25 M). This may be ascribed to the fast recombination of the photoinduced electron-hole pairs between the CB and VB of ZnIn 2 S 4 . Moreover, 0.2 wt.% was found to be the optimal BPQD loading, resulting in photocatalytic H 2 rates 1.3 times higher (3675.9 µmol·g −1 ·h −1 ) compared to pure ZnIn 2 S 4 . The total yield of H 2 increased significantly for the hybrid samples compared to pure ZnIn 2 S 4 . These results confirmed that decoration of BPQDs on ZnIn 2 S 4 improved the photocatalytic activity of the materials. A further increase in the BPQD loading above 0.2 wt.% resulted in lower H 2 production rates, probably because a fraction of the BPQDs acted as photo-recombination centers.
To confirm the positive effect of BPQDs in enhancing the photocatalytic H 2 production activity of ZIS, the photocatalytic H 2 production values of 0.2% BPQD-ZIS in this work were compared with the related literature, listed in Table 1. It can be clearly found that 0.2% BPQD-ZIS in this work showed superior photocatalytic performance to that of other quantum dots, implying that the incorporation BPQDs onto the ZIS surface may be a feasible route to improve its photocatalytic activity.  Recycling tests were carried out to evaluate the photostability of 0.2% BPQD-ZIS (Figure 7b). The yield of H 2 remained relatively stable for four consecutive runs, indicating the stability of the BPQD-ZIS sample during the photocatalytic H 2 production reaction.
The charge recombination properties of the as-prepared samples were further confirmed by PL measurements. Normally, the PL intensity is related to the recombination of holes and electrons. The lower PL intensity corresponds to the lower recombination of the photogenerated electron-hole pairs. The PL spectra of all as-prepared samples are shown in Figure 8. All samples showed a similar PL peak and the intensity of BPQD-ZIS hybrids obviously decreased compared with that of pure ZIS. The weakest peak intensity suggested the recombination of photogenerated electron-hole pairs was efficiently suppressed in the 0.2% BPQD-ZIS. However, when the BPQD loading mass was more than 0.2 wt.%, an evidently increased PL emission intensity occurred, which might be ascribed to the excess BPQDs acting as photorecombination centers, leading to lower H 2 production rates.
The transient photocurrent response and EIS were conducted to evaluate the charge separation and migration behavior. As shown in the Figure 9a, the pure ZIS exhibited a weak photocurrent response, owing to the fast photogenerated carrier recombination, while the photocurrent response of 0.2% BPQD-ZIS was much higher than that of the bare ZIS, which suggested that the charge recombination was remarkably inhibited. It can be clearly found in Figure 9b that the 0.2% BPQD-ZIS sample displayed a smaller semicircle than that of pure ZIS, suggesting that the charge transfer resistance can be efficiently reduced with the incorporation of BPQDs. Both the results of photocurrent and EIS experiments indicated that the incorporation of BPQDs onto the ZIS surface can facilitate charge carrier separation and transfer in the photocatalyst. The transient photocurrent response and EIS were conducted to evaluate the charge separation and migration behavior. As shown in the Figure 9a, the pure ZIS exhibited a weak photocurrent response, owing to the fast photogenerated carrier recombination, while the photocurrent response of 0.2% BPQD-ZIS was much higher than that of the bare ZIS, which suggested that the charge recombination was remarkably inhibited. It can be clearly found in Figure 9b that the 0.2% BPQD-ZIS sample displayed a smaller semicircle than that of pure ZIS, suggesting that the charge transfer resistance can be efficiently reduced with the incorporation of BPQDs. Both the results of photocurrent and EIS experiments indicated that the incorporation of BPQDs onto the ZIS surface can facilitate charge carrier separation and transfer in the photocatalyst.   The transient photocurrent response and EIS were conducted to evaluate the charge separation and migration behavior. As shown in the Figure 9a, the pure ZIS exhibited a weak photocurrent response, owing to the fast photogenerated carrier recombination, while the photocurrent response of 0.2% BPQD-ZIS was much higher than that of the bare ZIS, which suggested that the charge recombination was remarkably inhibited. It can be clearly found in Figure 9b that the 0.2% BPQD-ZIS sample displayed a smaller semicircle than that of pure ZIS, suggesting that the charge transfer resistance can be efficiently reduced with the incorporation of BPQDs. Both the results of photocurrent and EIS experiments indicated that the incorporation of BPQDs onto the ZIS surface can facilitate charge carrier separation and transfer in the photocatalyst. The Mott-Schottky (M-S) plots were conducted to determine the flat-band potential for pure ZIS and BPQDs. As seen from Figure 10, the positive slope of the M-S plots signified the type of ZIS and BPQDs as both n-type semiconductors, and, on the basis of the M-S equation, the flat-band potentials (Ufb) for ZIS and BPQDs were determined as −0.54 and −0.3 V vs. NHE (Normal Hydrogen Electrode), respectively. Normally, the bottom of the conduction band (CB) position of an n-type semiconductor is 0.1 V more negative than the flat-band potential [47]. Therefore, the conduction band potentials (CB) of ZIS and BPQDs are approximately −0.64 and −0.4 V. Meanwhile, combining the DRS results and the equation EVB = ECB + Eg, the values of the valence band potential (EVB) were determined as 1.65 and 2.28 V vs. NHE for ZIS and BPQDs, respectively. The Mott-Schottky (M-S) plots were conducted to determine the flat-band potential for pure ZIS and BPQDs. As seen from Figure 10, the positive slope of the M-S plots signified the type of ZIS and BPQDs as both n-type semiconductors, and, on the basis of the M-S equation, the flat-band potentials (U fb ) for ZIS and BPQDs were determined as −0.54 and −0.3 V vs. NHE (Normal Hydrogen Electrode), respectively. Normally, the bottom of the conduction band (CB) position of an n-type semiconductor is 0.1 V more negative than the flat-band potential [47]. Therefore, the conduction band potentials (CB) of ZIS and BPQDs are approximately −0.64 and −0.4 V. Meanwhile, combining the DRS results and the equation E VB = E CB + Eg, the values of the valence band potential (E VB ) were determined as 1.65 and 2.28 V vs. NHE for ZIS and BPQDs, respectively.

Mechanism behind the Enhanced Photocatalytic Performance of the BPQD-ZIS Hybrids
Based on above analysis, we propose a mechanism to explain the BPQD-ZIS photocatalytic H 2 evolution under light irradiation (Scheme 1). Under simulated sunlight irradiation, ZnIn 2 S 4 is photoexcited and generates electrons and holes. The electrons are transfer to the CB, while the holes remain in the VB. Typically, these electrons and holes are recombined rather than participating in the photocatalytic process. In the presence of BPQDs, the photo-generated electrons can be transferred to the surface of BPQDs quickly because of the lower energy level of BPQDs. Meanwhile, the photo-generated holes can migrate to the VB of ZnIn 2 S 4 , achieving efficient separation of charges. Thus, the electrons on the surface of BPQDs can participate in the H 2 evolution reaction, and the holes on the surface of ZnIn 2 S 4 can finish the oxidation process.

Conclusions
In summary, we, for the first time, decorated BPQDs on layered flower-like ZnIn 2 S 4 microspheres and used the resulting materials for photocatalytic H 2 generation. XRD, FTIR, UV-Vis absorption, and XPS were used to study the physicochemical properties of the samples. These characterization results confirmed that BPQDs were embedded into the nanosheets of the flower-like ZnIn 2 S 4 microspheres. The prepared BPQD-ZIS hybrids were quite stable during the photocatalytic H 2 generation in an aqueous solution. The enhanced photocatalytic properties of these materials can be attributed to the extended optical absorption and efficient charge separation and transport characteristics upon addition of BPQDs. The 0.2% BPQD-ZIS sample showed an optimum H 2 yield under simulated