Insights Into Highly Improved Solar-Driven Photocatalytic Oxygen Evolution Over Integrated Ag3PO4/MoS2 Heterostructures

Oxygen evolution has been considered as the rate-determining step in photocatalytic water splitting due to its sluggish four-electron half-reaction rate, the development of oxygen-evolving photocatalysts with well-defined morphologies and superior interfacial contact is highly important for achieving high-performance solar water splitting. Herein, we report the fabrication of Ag3PO4/MoS2 nanocomposites and, for the first time, their use in photocatalytic water splitting into oxygen under LED light illumination. Ag3PO4 nanoparticles were found to be anchored evenly on the surface of MoS2 nanosheets, confirming an efficient hybridization of two semiconductor materials. A maximum oxygen-generating rate of 201.6 μmol · L−1 · g−1 · h−1 was determined when 200 mg MoS2 nanosheets were incorporated into Ag3PO4 nanoparticles, which is around 5 times higher than that of bulk Ag3PO4. Obvious enhancements in light-harvesting property, as well as electron-hole separation and charge transportation are revealed by the combination of different characterizations. ESR analysis verified that more active oxygen-containing radicals generate over illuminated Ag3PO4/MoS2 composite photocatalysts rather than irradiated Ag3PO4. The improvement in oxygen evolution performance of Ag3PO4/MoS2 composite photocatalysts is ascribed to wide spectra response in the visible-light region, more efficient charge separation, and enhanced oxidation capacity in the valence band (VB). This study provides new insights into the design and development of novel composite photocatalytic materials for solar-to-fuel conversion.


INTRODUCTION
Inspired by natural photosynthesis, the construction of visible-light-responsive functional semiconducting materials for highly efficient photocatalytic water splitting and reduction of CO 2 has drawn considerable attention over the past few years (Maeda and Domen, 2010;Mikkelsen et al., 2010;Takanabe, 2017;Zheng et al., 2017). Especially, water splitting into hydrogen and oxygen has been intensively investigated due to the nature of clean and sustainable solar-to-fuel conversion. Compared with the hydrogen evolution reaction (HER), four-electron water oxidation process is more difficult to be fulfilled since a higher potential more than 1.23 eV is required (Kudo and Miseki, 2009), mostly an overpotential is also required. Thus, oxygen evolution is considered as the ratedetermining step in photocatalytic overall water splitting process. So far only few semiconductors such as WO 3 , BiVO 4 , have been employed as photocatalysts for oxygen production from water splitting (Xin et al., 2009;Wu et al., 2016;Zeng et al., 2017;He et al., 2018). Due to the limitations of band structures and light-harvesting properties in the visible light region, the utilization of sing-component semiconductors as catalysts for oxygen evolution has encountered serious difficulties, the design and development of novel composite materials for solar-driven photocatalytic water splitting are highly desirable.
Since the pioneer work of silver phosphate (Ag 3 PO 4 ) semiconductor for photocatalytic applications in 2010 (Yi et al., 2010). Many efforts have been devoted to synthesize Ag 3 PO 4 photocatalysts with different nanostructures and a variety of Ag 3 PO 4 -based composite photocatalytic materials for energy and environmental applications (Bi et al., 2012;Wang et al., 2012;Hu et al., 2013;Cao et al., 2017). Nanostructure engineering of pristine Ag 3 PO 4 and hybridization of Ag 3 PO 4 with other semiconductors have been proven to offer superior advantages in light absorption, electronhole separation and charge transportation, resulting in the enhancement in the photocatalytic activity (Yang et al., 2015b;Lv et al., 2016;Wang et al., 2017;Zhou et al., 2018). The key to synthesizing highly efficient Ag 3 PO 4 -based composite photocatalysts lies in screening promising candidates with matched band structures and constructing heterojunctions with optimal morphologies and interfaces, where favorable visible light utilization and tandem charge transfer pathway should be taken into consideration. The past decades have witnessed the use of molybdenum disulfide (MoS 2 ), a 2D lamellar material with excellent conductive property, as electrocatalysts and photocatalysts for applications in electrochemical and solar-tofuel conversion (Xiang et al., 2012;Dai et al., 2017;Yang et al., 2017). It was confirmed that the combination of MoS 2 with functional semiconductors enabled obvious enhancements in both photocatalytic and electrocatalytic hydrogen production activity (Sun et al., 2016;Zhang et al., 2016;Iqbal et al., 2017;Yuan et al., 2017). The hybridization of MoS 2 with Ag 3 PO 4 to produce MoS 2 /Ag 3 PO 4 composite materials has been primarily explored, however the application is restricted to the photodegradation of different kinds of organic pollutants (Wang L. et al., 2015;Wang P. F. et al., 2015;Gyawali and Lee, 2016;Li et al., 2016;Zhu et al., 2016;Wan et al., 2017), the employment of MoS 2 /Ag 3 PO 4 nanocomposites as photocatalysts for solar-lightdriven oxygen evolution from water splitting has not yet been explored.
Most recently, we reported the in-situ deposition of Ag 3 PO 4 on graphitic carbon nitride (g-C 3 N 4 ) nanostructures for highly efficient Z-scheme oxygen evolution from water splitting (Yang et al., 2015a,c;Cui et al., 2018;Tian et al., 2018). In consideration of the matched band structures of bulk Ag 3 PO 4 and MoS 2 materials for redox reactions, in this work, we demonstrate the hybridization of oxygen-producing photocatalyst Ag 3 PO 4 with few-layered, two-dimensional MoS 2 nanosheets, and the use of Ag 3 PO 4 /MoS 2 nanocomposites for photocatalytic water oxidation under LED illumination. Asprepared hybrid materials exhibit well-organized nanostructures, in which sheet-like MoS 2 materials provide sufficient active sites for the deposition of Ag 3 PO 4 nanoparticles. It is for the first time that oxygen evolution performance over the obtained Ag 3 PO 4 /MoS 2 composite photocatalysts has been evaluated, moreover, the effects of highly conductive MoS 2 materials on visible light utilization, electron-hole separation and water oxidation efficiency are systematically revealed.

EXPERIMENTAL Preparation
MoS 2 nanosheets were fabricated by the ultrasonic stripping of commercially available MoS 2 materials. In a typical synthesis of Ag 3 PO 4 /MoS 2 nanocomposites, different amounts of sheetlike MoS 2 (50, 100, 200, 300 mg) were added into 90 mL H 2 O, respectively, followed by the ultrasonic treatment for 3 h. Next, 30 mL of AgNO 3 (18 mmol, 3.06 g) aqueous solution was added dropwise into the MoS 2 suspension, and stirred for further 12 h. And then 30 mL of Na 3 PO 4 (6 mmol, 2.28 g) aqueous solution was added slowly into the above mixture, followed by continuous stirring for 3 h. After high-speed centrifugation, solid products were washed with deionized water and ethanol repeatedly, and dried at 60 • C for 12 h. The final products were collected and are denoted as AM-50, AM-100, AM-200, and AM-300.

Characterizations
X-ray diffraction (XRD) was measured using Cu Kα radiation with the 2θ range from 5 to 80 • at a scan rate of 5 • min −1 on D/MAX2500PC. Raman spectra were recorded by Thermo Scientific TM DXR spectrometer operating at 532 nm. X-ray photoelectron spectroscopy (XPS) was evaluated by Perkin-Elmer PHI 5000C. The field-emission scanning electron microscopy (FE-SEM) was performed on JSM-7001F. The Ultraviolet-visible diffuse reflectance spectrophotometer (UV-vis DRS) on UV2450 from 200 to 800 nm with BaSO 4 as reference standard. Photoluminescence (PL) emission measurements were carried out by a QuantaMaster TM 40 with an excitation wavelength of 420 nm. The electron spin resonance (ESR) measurements were recorded on a JES FA200 Spectrometer using the 5, 5-dimethyl-1-pyrroline-N-oxide (DMPO) as the radical capture reagent.

Photocatalytic Measurements
The efficiency of photocatalytic oxygen evolution was monitored by in-situ oxygen sensor in a sealed system connected with a circulation system. Before the measurement, oxygen-free and air-saturated water were used to calibrate the oxygen probe with temperature compensation. For the measurement of oxygen evolution, 0.3 g of the photocatalyst powder was added into AgNO 3 aqueous solution (100 mL, 10 g/L), followed by an ultrasonic treatment for 5 min.

RESULTS AND DISCUSSION
Photocatalytic oxygen evolution from water splitting over pure Ag 3 PO 4 and Ag 3 PO 4 /MoS 2 composites with different mass ratios were evaluated, and the results are presented in Figure 1. It can be observed ( Figure 1A) that the amount of evolved oxygen increases gradually when 50 and 100 mg MoS 2 were employed, the highest concentration of produced oxygen was recorded when the content of MoS 2 was increased to 200 mg. Further increase in the MoS 2 content from 200 to 300 mg in the Ag 3 PO 4 /MoS 2 composite resulted in the deterioration in the oxygen-generating performance. Notably all the composites showed improved oxygen-evolving performance than bulk Ag 3 PO 4 material. The oxygen-evolving rates of as-prepared samples under LED illumination are further quantified and shown in Figure 1B. When 200 mg MoS 2 was introduced to hybridize with Ag 3 PO 4 , an optimal oxygen-generating rate of 201.6 µmol · L −1 · g −1 · h −1 was determined, which is about 4.5 times higher than that of pure Ag 3 PO 4 . It can be concluded from the oxygen evolution performance that a proper addition of MoS 2 may promote the water oxidation efficiency, while the use of more than 200 mg of MoS 2 is found to show negative effects on the oxygen evolution performance. For simplicity, only the composite AM-200 with the best oxygen-producing efficiency is chosen for comparisons with two starting materials in the following sections.
Following the evaluation of photocatalytic oxygen evolution performance, phase structures of the AM-200 composite were confirmed by XRD patterns (Figure 2A). Diffraction peaks (black line) appearing in 14.32, 32.62, 33.44, 35,82, 39.48, 44,1, 49.72, 58.24, 60.32, and 72.72 • can be assigned to the (002), (100),  96, 29.78, 33.38, 36.66, 47.86, 52.76, 55.1, 57.34, 61.72, and 73.94   by Raman spectra (Figure 2B). In the spectrum of Ag 3 PO 4 , a sharp absorption peak at 908.5 cm −1 can be attributed to the motion of terminal oxygen bond vibration in phosphate chains. The peak at 1002.4 cm −1 is ascribed to the asymmetric stretching vibrations of O-P-O bonds in [PO 4 ] clusters. The broad peak located at 554.4 cm −1 arises from the asymmetric stretch of P-O-P bonds, while the peak centered at 406.1 cm −1 corresponds to the symmetric bending vibration modes related to [PO 4 ] clusters (Botelho et al., 2015). For pure MoS 2 , characteristic Raman shifts located at 375.8 and 402.1 cm −1 are assigned to the E 2g and A 1g modes, while the peak appearing at 446.2 cm −1 is suggested to come from the interaction of the longitudinal acoustic phonon and Raman inactive A 2u modes (Koroteev et al., 2011;Lukowski et al., 2013). All characteristic peaks of both Ag 3 PO 4 and MoS 2 were detected in Raman spectrum of the Ag 3 PO 4 /MoS 2 composite AM-200, indicating a complete hybridization of Ag 3 PO 4 with MoS 2 .
The morphologies of as-prepared samples were recorded by SEM images (Figure 3). From Figures 3A,B, as-synthesized Ag 3 PO 4 products present an irregular polyhedron structure with the size of about 1-3 µm, and a few of small particles are distributed around large particles. SEM images of ultrasonictreated MoS 2 material (Figures 3C,D) shows that stripped MoS 2 exhibit a thin sheet-like nanostructure. The stripped MoS 2 layer-like material has a certain accumulation and parts of MoS 2 materials have not been stripped into pieces. The Ag 3 PO 4 /MoS 2 composite AM-200 was synthesized by the in-situ deposition of Ag 3 PO 4 nanoparticles on the surface of MoS 2 nanosheets via electrostatically driven self-assembly. As displayed in Figures 3E,F Figure 4A) of as-prepared composite AM-200. The existence of C 1s peak is may due to the adventitious carbon on the surface of sample. In the high resolution spectrum of Ag 3d (Figure 4B), two peaks at 368.1 and 374.2 eV can be assigned to the Ag 3d 5/2 and Ag 3d 3/2 , respectively. The broad peak in the P 2p spectrum ( Figure 4C) located at 133.0 eV originates from the P 5+ in the Ag 3 PO 4 . The high resolution spectrum of Mo 3d is displayed in Figure 4D, the peaks at 226.7, 229.7, 232.9, and 236.2 eV can be ascribed to the S 2s, Mo 3d 5/2 , Mo 3d 3/2 , and Mo-O binding, respectively. Particularly, the first three binding energies indicated that S and Mo elements in the MoS 2 are found in the form of S 2− and Mo 4+ , respectively, and the last one might result from the exposed Mo atoms during the exfoliation process combining with the O of Ag 3 PO 4 (Wan et al., 2017). The S 2p XPS spectrum ( Figure 4E) can be divided into two peaks centered at 162.5 and 163.6 eV, respectively, corresponding to the S 2p 3/2 and S 2p 1/2 in the MoS 2 . The spectrum of O 1s (Figure 4F) can be fitted into two peaks located at 530.8 and 532.5 eV, originating from the O 2− in the Ag 3 PO 4 and the hydroxyl group, respectively.
It is well-known that the utilization of visible light is one of key factors affecting the activity of a photocatalyst, therefore, the light-harvesting properties of all samples were measured by UVvis DRS ranging from 200 to 800 nm and the absorption spectra are presented in Figure 5A. It can be seen that pure Ag 3 PO 4 has a clear absorption edge around 530 nm, and black MoS 2 material reveals full-spectrum absorption in the range of 200-800 nm. The light absorption intensity of Ag 3 PO 4 /MoS 2 composite AM-200 in the wavelength range of 500-800 nm was increased when a certain amount of MoS 2 (200 mg) were employed to hybridize with Ag 3 PO 4 , implying that the integration of MoS 2 with Ag 3 PO 4 favors a more efficient utilization of visible light. In addition to the light absorption, the separation of photoinduced electron-hole pairs is also believed to play a predominant role in determining the photocatalytic activity, thus the recombination of photogenerated charge carrier for the as-synthesized samples was analyzed by PL spectroscopy measurements. Figure 5B reveals Ag 3 PO 4 has a strong excitation peak around 630 nm, stemming from the recombination of electrons and holes. After the addition of MoS 2 , the Ag 3 PO 4 /MoS 2 composite AM-200 presented a similar position of excitation peak with Ag 3 PO 4 , and the PL emission intensity of AM-200 was weaker than those of pure MoS 2 and Ag 3 PO 4 , suggesting that the recombination efficiency of photoexcited charge carriers in the Ag 3 PO 4 /MoS 2 composite AM-200 has been effectively suppressed when the heterostructured composite was formed. A slower recombination rate of photogenerated electron-hole pairs boosts the enhancement in the photocatalytic performance.
To further determine the influence of the redox capacity of samples on the photocatalytic activity, as well as to investigate the mechanism behind the enhanced photocatalytic oxygen evolution from water splitting, the ESR measurement was carried out to confirm active radicals in-situ formed under light illumination. It can be observed in Figure 6A that no obvious peak was detected in dark. Under illumination, several strong peaks arising from DMPO-captured radicals can be detected in methanol dispersion for both pure Ag 3 PO 4 and the Ag 3 PO 4 /MoS 2 composite AM-200, typical peaks are assigned to the spin adducts (DMPO-O •− 2 ). Compared with signals derived from pure Ag 3 PO 4 , higher signal intensities of DMPO-captured superoxide radicals were observed in the methanol dispersion of AM-200. It is shown in Figure 6B that no radical signal was detected in dark. A typical intensity ratio of 1:2:2:1 was determined from the spin adducts in aqueous dispersions of both Ag 3 PO 4 and AM-200 under light irradiation, representing the generation of the spin adducts (DMPO-•OH). Similarly, the intensity of radical signal for AM-200 increased largely comparable to that of Ag 3 PO 4 .On the basis of the above results, it is concluded that higher intensities of both photo-induced O •− 2 and •OH were recorded when a proper amount of MoS 2 was employed.
Furthermore, the band edge positions of valence band (VB) and conduction band (CB) also have an important effect on the redox catalytic capability, which can be deduced by the following formula : Where E CB and E VB represent the CB and VB edge potentials, respectively; χ is the electro-negativity of the semiconductor, which is the geometric mean of the electro-negativities of the constituent atoms, and χ-values for Ag 3 PO 4 and MoS 2 are 5.96 and 5.32 eV (Wan et al., 2017), respectively. E e is about 4.5 eV, representing the free electron energy on the hydrogen scale. E g was the band gap energy of the semiconductor, and E gvalues for Ag 3 PO 4 and MoS 2 are about 2.45 and 1.9 eV (Yang et al., 2015c;Li et al., 2016), respectively. According to the calculation, the E VB -values of Ag 3 PO 4 and MoS 2 are about 2.69 and 1.77 eV, the top of which for both Ag 3 PO 4 and MoS 2 is more positive than the redox potential of O 2 /H 2 O (1.23 eV) (Xie et al., 2013), theoretically both two semiconductors are able to split water into oxygen. However, overpotential is generally required for practical water splitting, in this study, Ag 3 PO 4 acts as the oxygen-evolving catalyst for solar-driven water splitting due to its more positive potential higher than that for water oxidation. Subsequently E CB positions of Ag 3 PO 4 and MoS 2 are determined to be 0.24 and −0.13 eV, respectively. Therefore, the enhanced photocatalytic oxygen-generating performance over the Ag 3 PO 4 /MoS 2 composite photocatalyst could be explained as follows: first, the electrons in the VB of Ag 3 PO 4 could be initially excited into CB, and subsequently, may recombine with the holes in the VB of MoS 2 via a possible Z-scheme configuration. Thus, more efficient electron-hole separations and charge transporatation occur in the illuminated hybrid materials due to the existence of highly conductive MoS 2 sheets and possible Z-scheme pathway for electron transfer. The electronhole recombination on the surface of Ag 3 PO 4 can be suppressed, as a result, active holes left on the VB position of Ag 3 PO 4 may oxidize water into oxygen effectively, leading to highly efficient oxygen evolution performance over the Ag 3 PO 4 /MoS 2 composite photocatalysts.

CONCLUSIONS
In conclusion, effective Ag 3 PO 4 /MoS 2 composite photocatalysts were successfully fabricated by combining ion-exchange process and electrostatic assembly of Ag 3 PO 4 nanoparticles on the surface of MoS 2 nanosheets. The Ag 3 PO 4 /MoS 2 hybrid materials demonstrated superior interfacial contact and wide-spectrum light-harvesting property in the visible light region. When employed as the catalyst for photocatalytic water splitting, it exhibited highly improved oxygen evolution performance than bulk Ag 3 PO 4 under LED irradiation. The oxygen-evolving rate of the optimal Ag 3 PO 4 /MoS 2 composite (AM-200) is nearly five times faster than pure Ag 3 PO 4 . The combined characterizations and theoretical analysis on band structures suggest that the enhanced water oxidation efficiency is attributed to remarkable response to visible light, more efficient charge transportation and possibly specific Z-scheme pathway derived from matched band positions. The finding in this work offers a great opportunity in designing and synthesizing novel composite photocatalytic materials for applications in solar energy conversion, allowing us to develop an understanding of the fundamental mechanisms of Ag 3 PO 4 -based composite photocatalytic materials.

AUTHOR CONTRIBUTIONS
XY and QL: designed the project, guided the study, and polished the manuscript; XC, XX, and LT: conducted the experiments and characterized the samples; HT: revised the manuscript.