Mechanism study on the photocatalytic efficiency enhancement of MoS₂ modified Zn–AgIn₅S₈ quantum dots

Abstract

In this work, colloidal quantum dots (QDs) of environmentally friendly multinary semiconductor Zn–AgIn₅S₈ are used for heterostructure photocatalysts with hydrothermally deposited MoS₂ focusing on the mechanism study of charge transfer processes in the photocatalytic system. The influence of MoS₂ is systematically investigated on the structure, optical properties and photocatalytic activity of the Zn–AgIn₅S₈/MoS₂ nanocomposites, where the compromise of the catalytic activity and light shielding effect and increased stability is observed with increasing MoS₂. The optimized Zn–AgIn₅S₈/MoS₂-5% composite exhibits high photocatalytic activity for nearly complete degradation of Rhodamine B within 4 min under visible light irradiation. The composite photocatalysts also show good activity on degradation of other organic pollutants including a colorless antibiotic, tetracycline. Photocatalytic mechanism study proves that the main oxidizing species is the superoxide radical formed by the reaction of photogenerated electrons and oxygen molecules, for which MoS₂ mainly acts as a cocatalyst that promotes the electron transfer from the Zn–AgIn₅S₈ QDs. Time-resolved photoluminescence spectroscopy further reveals the increased charge separation by MoS₂ from the photoluminescence quenching and lifetime decrease of the QDs, especially that of the long-lived trap states. Moreover, more efficient charge transfer from the dyes is also observed with Zn–AgIn₅S₈/MoS₂, which is distinctively different from that with the Zn–AgIn₅S₈ QDs, indicating stronger interaction between the dye and photocatalyst with the introduction of MoS₂. Based on these results, a continuous charge flow mechanism is proposed, where the electron transfer can be promoted by MoS₂ not only from the QDs but also from the dyes that are all converted to superoxide radicals and contributed to the photocatalytic activity increase.

---

1. Introduction

Photocatalysis, as a widely studied sunlight-driven technique, has been considered as one of the most promising solutions for environment cleaning and renewable solar fuel production.^1–3 Generally speaking, an efficient photocatalyst requires sufficient sun light absorption, especially in the visible range, and effective separation and utilization of the photogenerated charge carriers. In the past decades, extensive efforts have been devoted to the development of efficient visible light photocatalysts.^4–6 Owing to the direct bandgap, high light absorption coefficient and low toxicity, multinary I–III–VI chalcogenide nanocrystals (NCs) (I = Cu, Ag; III = Al, In, Ga; VI = S, Se) have been comprehensively studied in the photocatalytic field,.^7–11 More importantly, the I–III–VI NCs hold a unique advantage for energy conversion applications that their composition can be varied in a large range while the lattice structure being maintained, which renders them excellent tunability of the band gap and band energy positions.^12–14 Photocatalytic degradation of organic dyes has been demonstrated for CuInS₂,¹⁵ AgInS₂ (ref. 16) and AgGaS₂ (ref. 17) nanostructures. Various alloyed and doped I–III–VI NCs have also been developed for photocatalysis by alloying within the I–III–VI materials¹⁸ or alloying with other materials, such as ZnS¹⁹ and CdS,²⁰ showing superior photocatalytic activity compared with the original counterparts. For example, AgIn₅S₈/AgInS₂,²¹ ZnS–AgInS₂,^22,23 AgInxGa_1−xS₂,¹⁷ and AgInZn_xS_x+2 (ref. 24) NC based photocatalysts have been developed with composition dependent photocatalytic activity indicating a promising future for photocatalytic applications by composition and structure design. However, these I–III–VI NCs still suffer from the common stability issue of sulfide materials and large effort on the photocatalyst structure design is needed for activity and stability improvement for practical applications.

A main consideration in photocatalyst design is the prompt separation and migration of the photogenerated electrons and holes, on which loading co-catalysts plays an important role.^15,25–27 Among these materials, MoS₂ is a 2D layered material with typical graphene-like structure that has attracted great attention in photocatalysis and electrocatalysis areas due to its low cost, high abundance and unique photoelectric properties.^28–31 MoS₂ has been combined with lots of semiconductors for photocatalytic applications, such as TiO₂, Bi₂MoO₆, BiVO₄, CdS and ZnIn₂S₄ etc., where it provides effective surface active sites that can prompt the surface catalytic reaction, inhibit recombination of photo-generated electrons and holes, and prevent the occurrence of counter reaction between active species and product.^20,32–35 Moreover, MoS₂ has a unique advantage for compositing with sulfide semiconductors that the shared S²⁻ could reduce the lattice mismatch and defect states at the interfaces. On the other hand, since the active sites are mainly the exposed edge sites of the 2D S–Mo–S planes, recently more and more work has been contributed to the size control of MoS₂ nanostructures towards high quality heterostructure photocatalysts with abundant active edge sites.²⁰ These research works suggest a promising future of sulfide semiconductors combined with MoS₂ for photocatalytic application. However, there has not been any report on size controlled MoS₂ combined with I–III–VI NCs and more profound investigation on the structure and interface property manipulation is still needed towards a clear understanding of the role of MoS₂ in the photocatalytic processes due to the complexity of the band energy alignment, structure, and deposition methods.

In the present study, series of Zn–AgIn₅S₈/MoS₂ nanocomposites were prepared by combining MoS₂ with monodispersed Zn–AgIn₅S₈ quantum dots (QDs) towards a more profound understanding of the role of MoS₂ in photocatalyst design. The effect of the MoS₂ content on the structure, optical properties and photocatalytic activity of the Zn–AgIn₅S₈/MoS₂ nanocomposites was systematically investigated towards better understanding of the crucial role of MoS₂ in the photocatalysis system. It was found that the visible-light-driven photocatalytic activity of Zn–AgIn₅S₈/MoS₂ nanocomposites were dramatically enhanced compared to pure Zn–AgIn₅S₈ QDs. The related photocatalytic mechanism study indicates that the superoxide radicals (˙O₂⁻) are the main active species with minor contribution from the holes. Steady state and time-resolved photoluminescence (PL) spectroscopies were used together for the first time to investigate the role of MoS₂ in the charge transfer properties of the Zn–AgIn₅S₈/MoS₂ photocatalysts in terms of PL quenching and lifetime change.

2. Experimental

2.1 Chemicals

All chemicals are analytical grade and used without further purification. Silver nitrate (AgNO₃), zinc acetate (Zn(OAc)₂·2H₂O), indium nitrate (In(NO₃)₃·4.5H₂O), thioacetamide (TAA), sodium hydroxide (NaOH), thiourea (CH₄N₂S), hexaammonium heptamolybdate ((NH₄)₆Mo₇O₂₄·4H₂O), ethylenediamine tetraacetic acid (EDTA) dipotassium salt (C₁₀H₁₄K₂N₂O₈·2H₂O), Rhodamine B (RhB), tetracycline (TC), methyl orange (MO), methyl red (MR) and absolute ethanol were purchased from Sinopharm Chemical Reagent Co. Ltd. L-Cysteine (HSCH₂CH(NH₂)CO₂H), 1,4-benzoquinone (BQ), isopropanol (IPA), and 5,5-dimethyl-1-pyrroline N-oxide (DMPO) were purchased from Aladdin (China).

2.2 Synthesis of Zn–AgIn₅S₈ QDs and Zn–AgIn₅S₈/MoS₂ composites

Pure Zn–AgIn₅S₈ QDs with Ag : In : Zn ratio of 2 : 10 : 5 were prepared by a one-pot hydrothermal method based on previous report with revisions.¹⁹ In a typical synthesis process, 0.34 mmol of AgNO₃, 1.7 mmol of In(NO₃)₃·4.5H₂O, and 0.85 mmol of Zn(OAc)₂·2H₂O were added into 5.5 mL of water and a clear solution was obtained after stirring. Then 2 mL of 2.5 M L-cysteine solution was added under vigorous stirring. The pH value of the mixture was adjusted to 8.5 using 1.0 M NaOH solution and then 3.25 mmol of thioacetamide was rapidly added. The solution was transferred to a 35 mL Teflon-lined stainless steel autoclave, sealed and heated at 110 °C for 4 h. After naturally cooling down to room temperature, the samples were collected by ethanol precipitation and centrifugation and were further washed with water/ethanol cycles for three times. The cleaned QDs were dispersed in water for further characterization and photocatalytic tests.

For preparation of Zn–AgIn₅S₈/MoS₂ composite materials, 300 mg of the purified Zn–AgIn₅S₈ QDs were added to 30 mL of deionized water containing appropriate amounts of (NH₄)₆Mo₇O₂₄·4H₂O and thiourea, where the molar ratio of Mo : S was kept at 7 : 30. Subsequently, the solution was transferred to a 50 mL Teflon-lined stainless steel autoclave and maintained at 200 °C for 8 h. The purification procedure of Zn–AgIn₅S₈/MoS₂ materials was the same as that of pure Zn–AgIn₅S₈ QDs. For comparison, a series of Zn–AgIn₅S₈/MoS₂ composites were prepared with different amount of MoS₂ ranging from 1 wt%, 3 wt%, 5 wt%, 10 wt%, 15 wt% to 30 wt%.

2.3 Characterizations

Structure of the samples was tested by X-ray diffraction (XRD, D8 ADVANCE, Bruker, Germany), using Cu-Kα radiation source (λ = 1.54056 Å) at a scanning rate of 4.0° per min. The transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) were also used to characterize the samples using a Tecnai G2 F30 S-Twin (FEI) electron microscope with an accelerating voltage of 200 kV. Raman spectra were recorded with an America ThermoFisher (DXR) Laser Raman Spectrometer with a 532 nm laser. The photoluminescence (PL) spectra of the samples were measured on a PerkinElmer LS 55 fluorescence spectrophotometer with an excitation wavelength of 450 nm. Time-resolved photoluminescence spectra were collected on a QuantaMaster™ 40 spectrometer (Photon Technology International, Inc.) with excitation wavelength of 481 nm. The electron spin resonance (ESR) spectra were collected on a Bruker EPR A300-10/12 spectrometer.

2.4 Photocatalytic activity evaluation and mechanism study

The photocatalytic activities of the Zn–AgIn₅S₈/MoS₂ composite materials were first evaluated via the photocatalytic degradation of RhB in aqueous solution under visible light irradiation. A 250 W Xe lamp with a 420 nm cutoff filter was used to provide visible light irradiation. In each experiment, 10 mg of photocatalyst was added into a 100 mL of the RhB solution (10 mg L⁻¹). Prior to irradiation, the suspensions were stirred in the dark for 60 min to reach the adsorption–desorption equilibrium. After light irradiation, 4 mL of the suspension was extracted at 2 min intervals and centrifuged to obtain a clear liquid that will be used for concentration measurement. In addition, to further explore its photocatalytic application, the Zn–AgIn₅S₈/MoS₂ composite photocatalysts were also investigated for degradation of different kinds of pollutants, including tetracycline (TC), methyl orange (MO) and methyl red (MR) with all the operations the same as that of RhB. Finally, the concentrations of RhB, TC, MO and TR were monitored using a TU-1810 UV-vis spectrometer at corresponding maximum absorption wavelengths of 553, 357, 463 and 410 nm, respectively. The dyes used in this work is mainly chosen in consideration of their hazard and the energy level alignment with the photocatalyst. With the rapid economic development, the pollution of wastewater from dyeing industry is the most dominant and tetracycline is an emerging antibiotic pollutant that is potentially dangerous for producing drug-resistant super bacteria. And RhB is further studied as the model dye for photodegradation mechanism because of its light absorption range and suitable energy levels that are suitable to investigate the interaction with our QDs based photocatalysts.

To probe the active species involved in the photocatalysis process, photocatalytic degradation of RhB has been studied with the introduction of different kinds of scavengers, where EDTA³⁶ as the hole scavenger, p-benzoquinone (BQ)³⁷ as the superoxide radical scavenger, and isopropyl alcohol (IPA)³⁸ as the hydroxyl radical scavenger. All other conditions were kept the same to the above photocatalytic experiments without scavengers. In addition, to further clarify the photocatalytic reaction mechanism, ESR technique was employed to detect the generated radicals in the photocatalytic system under visible light irradiation using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the radical trapping agent, which can trap the ˙O₂⁻ or ˙OH radicals and form DMPO–˙O₂⁻ or DMPO–˙OH complexes. Briefly, 2.5 mg of the Zn–AgIn₅S₈/MoS₂ sample was dispersed in methanol (for ˙O₂⁻) and then 30 μL of DMPO was added with ultrasonic dispersion for 3 min. ESR spectra of the trapped DMPO–˙O₂⁻ radicals were recorded before and after light irradiation.

3. Results and discussion

3.1 Structure and property characterization

XRD was used to characterize the phase of the as-prepared samples. As shown in Fig. 1a, the pure Zn–AgIn₅S₈ QDs basically match the characteristic peaks of cubic AgIn₅S₈ (JCPDS no. 25-1329).¹⁹ The peaks located at 27.3°, 46.6° and 55.0° can be assigned to the (311), (440) and (533) crystal planes of AgIn₅S₈, respectively. Owning to the small size of the particles, all of the diffraction peaks are relatively broad and weak. The top XRD pattern of MoS₂ synthesized with similar conditions can be indexed to hexagonal phase MoS₂ (JCPDS no. 73-1508) with the peaks of 29.5° and 58.5° assigned to the (101) and (110) crystal planes of MoS₂, respectively.²⁸ For the Zn–AgIn₅S₈/MoS₂ composites, the XRD patterns are similar to that of the initial QDs with no noticeable characteristic MoS₂ peak, even with the mass ratio of MoS₂ increases from 1 wt% to 30 wt%. This may be due to the small amount and the poor crystallinity of MoS₂ in the nanocomposites. Moreover, with respect to the invisibility of the MoS₂ diffraction peaks in the heterojunctions, the highly similar XRD patterns of the Zn–AgIn₅S₈/MoS₂ composites indicate that the low content of MoS₂ did not bring any influence on the crystal structure of the Zn–AgIn₅S₈ QDs.

Fig. 1 (a) XRD patterns of the pure Zn–AgIn₅S₈ QDs, MoS₂ and Zn–AgIn₅S₈/MoS₂ composites with different MoS₂ mass ratios of 1 wt%, 5 wt%, 10 wt%, 15 wt%, 30 wt%. (b) Raman spectra of the Zn–AgIn₅S₈ quantum dots, pure MoS₂ nanosheets and Zn–AgIn₅S₈/MoS₂ composites with different MoS₂ mass ratios of 1 wt%, 5 wt%, 10 wt%, 30 wt%.

Raman spectroscopy, as a more sensitive technique, was employed to further investigate the structure evolution of the Zn–AgIn₅S₈/MoS₂ nanocomposites in comparison with the Zn–AgIn₅S₈ QDs and pure MoS₂ nanosheets. As shown in Fig. 1b, the strong broad peaks around 318 (for Zn–AgIn₅S₈ QDs), 327 (for Zn–AgIn₅S₈/MoS₂-1%) and 343 cm⁻¹ (for Zn–AgIn₅S₈/MoS₂-5%) are due to the PL signal of the Zn–AgIn₅S₈ QDs excited by the 532 nm green laser of the Raman spectrometer. Different from pure QDs, a tiny peak at 372 cm⁻¹ was observed with 1 wt% and 5 wt% MoS₂ that can be assigned to the E_2g mode of MoS₂. With the increase of MoS₂ content, the PL peak gradually disappeared and the two Raman peaks at approximately 372 cm⁻¹ and 402 cm⁻¹ increased dramatically, which are assigned to the E_2g and A_1g modes of MoS₂, respectively. There are no pronounced peak shift of the E_2g and A_1g modes compared to those of the pure MoS₂ nanosheets.³¹ It is worth to mention that until MoS₂ content was higher than 5%, the E_2g and A_1g modes of MoS₂ fully appeared in the composites. The decrease of the PL peak of the QDs and increase of the E_2g and A_1g modes of MoS₂ suggest the strong interaction between the two components and the dynamic growing process with increase of MoS₂ ratio, which was further verified in the optical property studies presented below.

The morphology change of the Zn–AgIn₅S₈/MoS₂ composites with different MoS₂ content was investigated by TEM to study the hybridization process between the Zn–AgIn₅S₈ QDs and MoS₂. As shown in Fig. 2a, the quaternary Zn–AgIn₅S₈ QDs prepared using a low temperature hydrothermal method show sphere-shaped morphology with average diameter of 4.3 ± 0.8 nm, exhibiting relatively uniform size with a narrow size distribution. Fig. 2b and c show the TEM images of the samples with 5 wt% and 30 wt% of MoS₂, respectively. It can be seen that the pure Zn–AgIn₅S₈ QDs cover the surface of the MoS₂ slices. With the increasing amount of MoS₂, the nanocomposites showed size increase and the MoS₂ nanosheets became more and more condensed as indicated by the characteristic stacked sheet structure of 2D layered materials (Fig. 2c). At the same time, the MoS₂ slices also gradually coagulate together (Fig. 2c) instead of individual quantum size particles with increasing MoS₂ content. As illustrated in Fig. 2d and e, the MoS₂ nanosheets and Zn–AgIn₅S₈ QDs were strongly linked to each other, where the 2D sheet-like morphology of MoS₂ with length over 50 nm already started to form. The enlarged image of Zn–AgIn₅S₈/MoS₂-5% demonstrated that the Zn–AgIn₅S₈ QDs were attached to the MoS₂ slices and an intimate interfaces were well formed between Zn–AgIn₅S₈ and MoS₂ owning to the shared sulfide ions at the interface, which would favor the charge carrier transfer. The HRTEM images in Fig. 2d and e clearly presented the lattice spacing of 0.63 nm, which could be assigned as the (002) plane of MoS₂,²⁸ while the fringes with a lattice spacing of 0.33 nm correspond to the (311) plane of the Zn–AgIn₅S₈ QDs.¹⁹

Fig. 2 TEM images of the Zn–AgIn₅S₈/MoS₂ samples with different mass ratios of MoS₂: (a) 0 wt% (Zn–AgIn₅S₈ QDs), (b) 5 wt%, (c) 30 wt% and (d and e) HRTEM images of AgIn₅S₈/MoS₂-5%.

The optical property of Zn–AgIn₅S₈/MoS₂ nanocomposites with increasing MoS₂ ratio from 0 wt% to 30 wt% was examined using UV-vis absorption spectroscopy. As shown in Fig. 3, the absorption edge of the pure Zn–AgIn₅S₈ QDs was estimated to be ∼550 nm, while the Zn–AgIn₅S₈/MoS₂ composite samples showed absorption edges of 600–650 nm, which may be ascribed to the increased visible light absorption caused by MoS₂. Compared with pure Zn–AgIn₅S₈ QDs, the visible light absorption was improved dramatically with the increase of MoS₂ content with colors changing from bright orange to dark red (lower inset of Fig. 3). The sample color change is also partly brought by increased light scattering with particle growth, changing from the clear orange solution (pure QDs) to the turbid orange state (1% MoS₂) and further to dark red (5% MoS₂). When the MoS₂ content is higher than 5 wt%, further increase of light absorption was observed in the whole spectral range, which can be ascribed to the increased light scattering of the samples partly due to the size increase of composite particles. The UV-vis spectra show that the introduction of MoS₂ enhanced the light absorption dramatically, though it may not necessarily contribute to the improvement of photocatalytic activity. In addition, the band gap (E_g) of pure Zn–AgIn₅S₈ QDs composites was estimated to be 2.28 eV from the plot of (αhν)² vs. hν based on the UV/vis absorption data, as shown in the upper inset in Fig. 3.

Fig. 3 UV/vis absorption spectra of the Zn–AgIn₅S₈ QDs and Zn–AgIn₅S₈/MoS₂ nanocomposites with different MoS₂ content: 1 wt%, 5 wt%, 10 wt%, 15 wt%, and 30 wt%. The inset figures are the curves of (αhν)² vs. hν of Zn–AgIn₅S₈ QDs (upper) for band gap estimation and the photographs (lower) of the Zn–AgIn₅S₈ QDs and Zn–AgIn₅S₈/MoS₂ composites with 1 wt% and 5 wt% of MoS₂ dispersed in water.

Based on the above results, we propose a possible growing mechanism for the Zn–AgIn₅S₈/MoS₂ nanocomposites (Scheme 1), where the amount of MoS₂ has a major influence on both of the morphology and the optical properties. With lower MoS₂ ratio (1 wt% and 5 wt%), the small MoS₂ particles deposit on the surface of the QDs and the resulted Zn–AgIn₅S₈/MoS₂ composite particles remain good colloidal stability along with a color change from orange to dark red (Fig. 3). Further increase of MoS₂ results in the size increase of the composite particles due to the growth of the MoS₂ 2D slices and also brings high light absorption in a wide range and dark color. The influence of these results was further investigated in photocatalytic applications.

Scheme 1 Schematic illustration of the in situ growth process of MoS₂ on the Zn–AgIn₅S₈ QDs for the construction of heterostructure photocatalysts.

3.2 Photocatalytic activity

RhB was used as the probing molecule to explore the photocatalytic degradation activity of the Zn–AgIn₅S₈/MoS₂ nanocomposites under visible-light irradiation. Fig. 4a shows the photocatalytic activities of the Zn–AgIn₅S₈/MoS₂ nanocomposites with different MoS₂ contents as well as the pure Zn–AgIn₅S₈ QDs and pure MoS₂ nanosheets under visible-light irradiation. It can be found that RhB is only slightly degraded in the presence of MoS₂ only or in the absence of various kinds of the prepared samples, indicating that the photocatalytic activity by MoS₂ itself can be ignored and there is no noticeable decolorization of the dye without photocatalysts. Pure Zn–AgIn₅S₈ QDs showed relatively high photocatalytic activity for degradation of RhB (88%) under visible light within 8 min, which may be due to the small size that provides high surface area for the photocatalytic reaction to occur. Along with the introduction of different contents of MoS₂, the photocatalytic performance of Zn–AgIn₅S₈/MoS₂ heterojunction could be enhanced. The Zn–AgIn₅S₈/MoS₂-5% composite shows the highest activity that RhB can be completely photodegraded in 4 min, which is much better than that of pure Zn–AgIn₅S₈ QDs. This degradation efficiency is much higher than that of previously reported AgIn₅S₈ powders (89.4% in 4 h for MO),³⁹ monodispersed AgIn₅S₈ microspheres (98% in 20 min for MO) or ZnIn₂S₄/AgIn₅S₈ heteromicrospheres (99% in 50 min for RhB).⁴⁰ The boosted photocatalytic activity may be attributed to the high surface area of the small size QD-based nanocomposites and the intimate interface between Zn–AgIn₅S₈ and MoS₂ due to the shared S²⁻ anions that can facilitate the interfacial charge transfer.^20,32

Fig. 4 Photocatalytic activity (a) and the pseudo-first-order reaction kinetics (b) for degradation of RhB using the Zn–AgIn₅S₈/MoS₂ composites with different amount of MoS₂ under visible light irradiation.

Moreover, the photocatalytic degradation kinetics results indicate that the reaction rate fit well with the pseudo-first order correlation as shown in Fig. 4b. It can be seen that the Zn–AgIn₅S₈/MoS₂-5% sample exhibits the highest degradation rate, which is about 3.35 times higher than that of pure Zn–AgIn₅S₈ QDs. With increasing MoS₂ content, the photocatalytic activity of the Zn–AgIn₅S₈/MoS₂ composite materials was enhanced at first and then weakened. When the content of MoS₂ is 1 wt% and 3 wt%, the photocatalytic activity has gradually enhanced comparing with the pure Zn–AgIn₅S₈ QDs. However, when MoS₂ loading exceeds 5 wt%, the photocatalytic activity of the 10 wt%, 15 wt% and 30 wt% samples started to decrease gradually compared to the 5 wt% sample. The poorer photocatalytic activity of the composites than pure QDs is only observed with high ratio of MoS₂ (15 wt% and 30 wt%). These results indicate that the optimal loading amount of Zn–AgIn₅S₈/MoS₂ sample was 5 wt% in terms of photocatalytic activity, which suggests that MoS₂ acts as a cocatalyst instead of light absorber and works best at relatively low loading amount. There are two contrast impacts with the introduction of MoS₂: the increased charge separation (positive effect) and the increased light shielding (negative effect). The decrease of photocatalytic activity at high loading amount (10–30 wt%) may be due the strong light absorption of MoS₂ that shielded the light absorption of the QDs similar to previously reported work.⁴¹ This observation indicates that the loading amount of MoS₂ has to be compromised considering the catalytic activity enhancement and the light shielding effect.

In order to evaluate the stability and reusability of the Zn–AgIn₅S₈/MoS₂ composite photocatalysts, cycle runs of photocatalytic degradation of RhB were carried out under visible-light irradiation. As shown in Fig. 5a, the efficiency of Zn–AgIn₅S₈/MoS₂-5% decreases from 100% in the 1st run to 65% in the 3rd run, which may be due to the incomplete recycle of the small sized photocatalyst. On the other hand, the photocatalytic activity of Zn–AgIn₅S₈/MoS₂-30% decreases from 100% in the 1st run to 90% in the 3rd run (Fig. 5b), e.g. the stability of the Zn–AgIn₅S₈/MoS₂ photocatalysts was enhanced with the increase of MoS₂. A plausible explanation is that the presence of extra MoS₂ contributed to the efficient separation of the photo-generated electrons and holes. The photocatalytic activity of the AgIn₅S₈/MoS₂-30% sample has a little deactivation (<10%) after three successive cycles for the degradation of RhB under visible-light irradiation, indicating that our photocatalyst possesses relatively high stability for practical applications.³²

Fig. 5 Cycle runs of photocatalytic degradation of RhB under visible light irradiation over the Zn–AgIn₅S₈/MoS₂ samples with 5 wt% (a) and 30 wt% (b) of MoS₂.

To further investigate their photocatalytic applications, the Zn–AgIn₅S₈/MoS₂-5% sample was also evaluated for degradation of other dyes (MO and MR) and an antibiotic (TC). Fig. 6 shows the photocatalytic performance on the degradation of TC, MO and MR under visible light. It was found that the Zn–AgIn₅S₈/MoS₂-5% nanocomposite could also efficiently decompose these organic pollutants, where TC could be degraded (74%) in 4 min and MO and MR could be completely degraded within 12 and 8 min. Hence, it is implied that our Zn–AgIn₅S₈/MoS₂ photocatalysts display excellent photocatalytic activity in the visible-light range and has great potential in environment purification.

Fig. 6 Photocatalytic activity of the Zn–AgIn₅S₈/MoS₂-5% nanocomposite for the degradation of TC, MO and MR under visible light irradiation.

3.3 Photocatalytic mechanism study

Active species. To better understand the photocatalysis mechanism, trapping experiments of the active species were carried out during photocatalytic degradation of RhB. As shown in Fig. 7a, for Zn–AgIn₅S₈/MoS₂-5%, the photocatalytic degradation of RhB shows no obvious change (red curve) with IPA (quencher of ˙OH) added compared with no quencher, which indicates that ˙OH is not the main reactive species. When EDTA (quencher of h⁺) was added into the solution, the degradation of RhB slowed down slightly, but was still completed within 12 min, which indicates that the photogenerated holes play a minor role in the photocatalytic system. In contrast, addition of BQ (˙O₂⁻ scavenger) caused complete quenching of photodegradation, implying that ˙O₂⁻ is the major reactive species and plays a predominant role in the photocatalytic process. Therefore, it is clearly illustrated that ˙O₂⁻ from the reaction of photogenerated electrons and O₂ are the main oxidative species, while the photogenerated h⁺ also contributed in certain extent for the degradation of RhB in the Zn–AgIn₅S₈/MoS₂ composite photocatalyst system.

Fig. 7 (a) Trapping experiment of the active species during photocatalytic degradation of RhB over Zn–AgIn₅S₈/MoS₂-5% nanocomposite alone and with the addition of EDTA (quencher of h⁺), BQ (quencher of ˙O₂⁻), and IPA (quencher of ˙OH) under visible light irradiation. (b) ESR spectra of the superoxide radical adducts trapped by DMPO in Zn–AgIn₅S₈/MoS₂-5% methanol dispersion under visible light irradiation.

In order to further reveal the reactive oxidation species formed over the Zn–AgIn₅S₈/MoS₂ materials, ESR spectroscopy was performed after light irradiation. It is known that DMPO is generally used as a radical trapping agent forming DMPO–˙O₂⁻ or DMPO–˙OH.^37,38 As shown in Fig. 7b, the twelve characteristic peaks of DMPO–˙O₂⁻ can be clearly observed in the methanol dispersion of Zn–AgIn₅S₈/MoS₂-5% nanocomposite after visible light irradiation, indicating that the ˙O₂⁻ radicals were produced from the reaction of photogenerated electrons and O₂ molecules in the photocatalytic systems. Combined with the trapping experiment (Fig. 7a), it can be inferred that ˙O₂⁻ played a major role in the photocatalytic system.

Charge transfer and reaction processes. On the basis of the above results, a probable mechanism is proposed for the enhanced photocatalytic activity of Zn–AgIn₅S₈/MoS₂ nanocomposites. It was reported that AgIn₅S₈ has direct bandgaps of 1.7–1.8 eV,⁴⁷ as one of the most potential candidates for the visible-light-driven photocatalytic applications. Here the effect of Zn doping of AgIn₅S₈ QDs in the whole system can be divided into three parts: changing absorption threshold, shifting the band energy levels and reducing the defect states. And the crucial and positive influence is increasing the conduction band and reducing the defect states according to the literatures. As illustrated in Scheme 2, the Zn–AgIn₅S₈ QDs can be excited under visible-light illumination, leading to the generation of electron–hole pairs. The photoexcited electrons can be transferred from CB of the Zn–AgIn₅S₈ QDs to MoS₂ due to the proper band energy alignment, and the separated electrons accumulated in the CB of MoS₂ can reduce the adsorbed O₂ to form ˙O₂⁻ radicals, which is a powerful oxidative species that can react with organic pollutants and convert them to small molecules. Meanwhile, the photogenerated holes left in the VB of Zn–AgIn₅S₈ can directly oxidize organic pollutants adsorbed on the surface of the photocatalyst, though it only accounts for a small portion of the photodegradation. Here MoS₂ is more like a cocatalyst that can enhance the charge separation and accept the excited electrons in the photocatalyst system, which results in the efficient production of ˙O₂⁻, the main oxidative species. To better explain the charge separation process in the Zn–AgIn₅S₈/MoS₂ composites, the band edge potentials of the Zn–AgIn₅S₈ QDs were estimated by the following equations:

E_CB = X − E^e − 0.5E_g (1)

E_VB = E_CB + E_g (2)

where E_CB and E_VB are the conduction and valence band edge potentials, X is the electronegativity of the semiconductor from the geometric mean of the electronegativity of the constituent atoms, E^e is the energy of free electrons on the hydrogen scale (∼4.5 eV), and E_g is the band gap of the semiconductor (2.28 eV here for the Zn–AgIn₅S₈ QDs as estimated from Fig. 3). The Zn effect on QD E_CB and E_VB energy levels is mainly determined by the electronegativity of the Zn–AgIn₅S₈ semiconductor. It is the geometric mean value of the absolute electronegativity of each atom of the compound. According to the initial ratio of each element, Zn doping of AgIn₅S₈ QDs forms an alloy, such as AgIn₅S₈–(ZnS)₅. After calculating, X approximately equals to 4.9 eV. Combined with the above equation, the CB and VB potentials are about −0.74 and 1.54 V for the Zn–AgIn₅S₈ QDs. It was reported that the CB and VB potentials of AgIn₅S₈ are −0.66 and 1.10 eV, while the electronegativity of AgIn₅S₈ is 4.72 eV.⁴⁸ Our results indicate that Zn doping of AgIn₅S₈ QDs could increase the conduction band from −0.66 eV to −0.74 eV, which increases the reducing ability of the photogenerated electrons. When the Zn–AgIn₅S₈ QDs is excited, electrons can be efficiently transferred to the MoS₂ cocatalyst and reduce the surface-adsorbed O₂ into ˙O₂⁻ (−0.046 V vs. NHE), inducing the formation of the active oxidizing species.^32,36

Scheme 2 The proposed photocatalytic mechanism of the Zn–AgIn₅S₈/MoS₂ nanocomposites for degradation of organic pollutants under visible light irradiation: the MoS₂ enhanced electron injection from the QDs, the subsequently increased interaction between the dyes and photocatalysts, and the formation of superoxide radicals.

On the other hand, it was reported that the strong interaction between the dyes and photocatalysts plays an important role in lots of photocatalytic processes.⁴ In this Zn–AgIn₅S₈/MoS₂ system, we propose that the dye/photocatalyst interaction can be effectively enhanced with the existence of MoS₂ since the photogenerated electrons in the QDs can be promptly transferred. In the photocatalytic reaction, electrons from RhB-LUMO (−1.42 V vs. NHE)⁴ can be injected into the Zn–AgIn₅S₈ QDs conduction band (−0.74 V vs. NHE). The matching band energy levels (RhB, Zn–AgIn₅S₈ and MoS₂) are shown in Scheme 2 and this synergistic effect are favorable for photogenerated electron–hole separation and transfer at both interfaces. To better demonstrate the effect of MoS₂ in the photocatalytic system, the charge separation processes were further investigated by PL quenching and lifetime measurements for both of the QDs and RhB as discussed below.

Enhanced charge transfer between the two components by MoS₂. In order to explore the charge separation and recombination processes in the prepared photocatalysts, PL spectroscopy was used to analyze the photocatalytic mechanism. As shown in Fig. 8a, the Zn–AgIn₅S₈ QDs show a strong PL peak centered at 646 nm with excitation of 450 nm. With the increase of MoS₂ content, PL intensity of the Zn–AgIn₅S₈/MoS₂ composites decreased significantly from 468 to 106 (for 1 wt%) and 14 (for 5 wt%) and reached almost zero when the loading amount exceeded 5 wt% (data not presented). This phenomenon was also observed in the Raman spectra as shown in Fig. 1b. The PL intensity decrease implies the enhanced charge separation and eliminated radiative recombination of the photo-generated electrons and holes in the composites. This result clearly indicates that MoS₂ has a major influence on the charge separation and subsequently photocatalytic activities of Zn–AgIn₅S₈/MoS₂ samples. The Zn–AgIn₅S₈/MoS₂-5% composite shows the highest photocatalytic performance compared with the other samples, which reveals efficient transfer of photo-excited electrons from Zn–AgIn₅S₈ QDs to MoS₂ and the suppressed radiative recombination of the photo-generated charge carriers.

Fig. 8 (a) Evolution of the PL spectra and (b) corresponding PL decay curves, and bi-exponential fitting curves of the Zn–AgIn₅S₈ QDs and Zn–AgIn₅S₈/MoS₂ samples with MoS₂ mass ratios of 1 wt% and 5 wt%.

To further probe the mechanism of electron–hole recombination in the Zn–AgIn₅S₈/MoS₂ samples, time-resolved PL spectroscopy was used to assess the PL lifetimes (Fig. 8b). The PL decay curve of the prepared sample can be well fitted with the biexponential function:

I(t) = A₁ exp(−t/τ₁) + A₂ exp(−t/τ₂) (3)

τ_ave = (A₁τ₁ + A₂τ₂)/(A₁ + A₂) (4)

where I(t) is the PL intensity at delay time of t, A₁ and A₂ are the relative weights of the decay components, τ₁ and τ₂ represent the lifetimes of two decay pathways. In comparison with Zn–AgIn₅S₈ QDs (τ_ave = 156.05 ns), the PL lifetime of Zn–AgIn₅S₈/MoS₂ composites with 1 wt% and 5 wt% MoS₂ (τ_ave = 143.6 and 14.85 ns, respectively) decrease rapidly with increasing MoS₂. The addition of MoS₂ results in the efficient PL quenching and lifetime decrease. It was reported that the fast decay component (τ₁) can be attributed to the surface trap state recombination, while the slowly decay component (τ₂) can be ascribed to deep donor–acceptor recombination process.^19,42,43 When loaded with MoS₂, lifetime of both the surface states and deep trap states were dramatically decreased (Table 1) indicating the effective quenching of the photogenerated carriers, which can be attributed to the efficient electron transfer from surface states or deep trap states near CB of the Zn–AgIn₅S₈ QDs to MoS₂ that is competing with the radiative recombination process. Moreover, a more efficient quenching of the long lived deep trap states (A₂ changing from 63.9% to 61.5% and 9.8% with increasing MoS₂) than the short lived surface trap states (A₁ in Table 1.) was observed, which suggests that introduction of MoS₂ makes the electron transfer and ˙O₂⁻ formation processes more favorable than the electron transfer to the deep trap states within the QDs. The electrons transferred to MoS₂ will further react with O₂ to form ˙O₂⁻ radicals for photocatalytic reactions.³³ These time-resolved PL results are in good agreement with the decrease of PL intensity and increase of photocatalytic activity, which further demonstrates that MoS₂ acts as a cocatalyst, rather than a light absorber, and mainly contributes to the increased charge separation during the photocatalytic process.

Table 1 PL decay parameters of the Zn–AgIn₅S₈ QDs and Zn–AgIn₅S₈/MoS₂ composites

	A1/%	τ1/ns	A2/%	τ2/ns	τave/ns
Zn–AgIn5S8 QDs	36.1	9.95	63.9	238.75	156.05
Zn–AgIn5S8/MoS2-1%	38.5	27.09	61.5	216.45	143.6
Zn–AgIn5S8/MoS2-5%	90.2	6.08	9.8	95.54	14.85

---

Enhanced interaction between the dye and photocatalysts by MoS₂. It has been reported that the interaction between dyes and semiconductor QDs, such as CdSe and CdS, plays an important role in the charge transfer and photoelectric conversion processes.^44–46 To better understand the role of MoS₂ in our photocatalytic system, more profound study was performed in terms of PL quenching and lifetime change of RhB upon the addition of the Zn–AgIn₅S₈ QDs and the Zn–AgIn₅S₈/MoS₂-5% composite, which showed distinctively different behaviors. As shown in Fig. 9a, when quantitative (20 μL) Zn–AgIn₅S₈ QDs solution (3 mg mL⁻¹) was added to RhB solution continually, the PL peak intensity at 586 nm of RhB gradually decreased, while PL of the Zn–AgIn₅S₈ QDs (around 600–750 nm) increased slowly. It was found that the PL intensity was reduced from 730 for RhB to 385 for 1 : 10 mixed solution (the condition of the photocatalytic reaction), but the quenching effect was slowed down with further addition of the QDs. In sharp contrast, nearly 100 times PL quenching of RhB by addition of Zn–AgIn₅S₈/MoS₂-5% was observed, from 730 to below 10 for RhB : Zn–AgIn₅S₈/MoS₂ = 1 : 10 (Fig. 9b), e.g. nearly complete PL quenching in the actual photocatalytic reaction condition. Fig. 9c shows the relationship of the PL intensity vs. ratio of RhB/photocatalysts, which clearly indicates the much more efficient PL quenching by the Zn–AgIn₅S₈/MoS₂ composite than the QDs. We propose that this enhanced interaction between RhB and Zn–AgIn₅S₈/MoS₂ photocatalysts can be attributed to the MoS₂ assisted electron transfer processes illustrated in Scheme 2.

Fig. 9 PL spectra evolution of the RhB solution with continuous addition of the Zn–AgIn₅S₈ QDs (a) and Zn–AgIn₅S₈/MoS₂-5% nanocomposites (b); (c) the corresponding PL intensity change vs. mass ratio of the dye/photocatalysts (the arrow position corresponds to the actual ratio for the photocatalytic reaction); (d) PL decay curves, corresponding single-exponential fitting curves and lifetime data of RhB, RhB with Zn–AgIn₅S₈ QDs and RhB with Zn–AgIn₅S₈/MoS₂-5%, respectively.

Further exploration by the time-resolved PL spectra of RhB with and without MoS₂ gives more profound understanding of the charge separation and recombination processes. As shown in Fig. 9d, for RhB/Zn–AgIn₅S₈ (1 : 4) mixture, the PL intensity at 586 nm was decreased, but the lifetime remains almost unchanged (2.32 ± 0.04 ns) compared to that of the original RhB solution (2.31 ± 0.04 ns), which indicates that this is an energy transfer process rather than a charge transfer process. However, the PL lifetime of RhB was decreased from 2.31 ns to 1.91 ns when Zn–AgIn₅S₈/MoS₂-5% was added, indicating a change of the excited state decay kinetics and the effective charge transfer from RhB to the Zn–AgIn₅S₈/MoS₂-5% photocatalyst.^44,45

4. Conclusions

In summary, series of Zn–AgIn₅S₈/MoS₂ nanocomposite photocatalysts were prepared by a facile hydrothermal method. The amount of MoS₂ played a key role on the structure and optical property of the Zn–AgIn₅S₈/MoS₂ nanocomposites and the compromise of catalytic activity enhancement and light shielding effect was observed with continuous growth of MoS₂. Photocatalytic mechanism study indicates that the main oxidation species is superoxide radicals produced from the photogenerated electrons and oxygen molecules. The high photocatalytic degradation activity of organic pollutants by these composite photocatalysts can be attributed to the small size and the intimate interfacial contact that promotes the electron–hole separation and transfer. Moreover, stead state and time-resolved PL spectroscopy studies suggest that MoS₂ mainly acts as a cocatalyst that promotes the electron transfer from the Zn–AgIn₅S₈ QDs. By further PL quenching and lifetime measurements of RhB upon the addition of Zn–AgIn₅S₈/MoS₂ composites, it was found that MoS₂ has a major influence on the interaction between the dye molecules and photocatalysts, for which a continuous charge flow mechanism was proposed with the introduction of MoS₂. These findings provide an interesting view and useful guidance for the use of MoS₂ in photocatalyst design in the future.

Acknowledgements

The authors would like to acknowledge the supports from the National Natural Science Foundation of China (21501072 and 21522603), the Chinese-German Cooperation Research Project (GZ1091), the Distinguished Professor Program of Jiangsu Province, the Natural Science Foundation of Jiangsu Province (BK20150489), the “Innovative and Entrepreneurial Doctor” Program of Jiangsu Province, the China Postdoctoral Science Foundation (2016M590419), the Jiangsu Province Postdoctoral Foundation (1501027A), and the Start Funding of Jiangsu University (15JDG011 and 15JDG027).

References

1. F. Fresno, R. Portela, S. Suarez and J. M. Coronado, J. Mater. Chem. A, 2014, 2, 2863–2884.
2. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520–7535.
3. J. R. Ran, J. Zhang, J. G. Yu, M. Jaroniec and S. Z. Qiao, Chem. Soc. Rev., 2014, 43, 7787–7812.
4. O. F. Lopes, E. C. Paris and C. Ribeiro, Appl. Catal., B, 2014, 144, 800–808.
5. H. Tong, S. Ouyang, Y. Bi, N. Umezawa, M. Oshikiri and J. Ye, Adv. Mater., 2012, 24, 229–251.
6. N. Razgoniaeva, P. Moroz, S. Lambright and M. Zamkov, J. Phys. Chem. Lett., 2015, 6, 4352–4359.
7. T. A. Kandiel, D. H. Anjum and K. Takanabe, ChemSusChem, 2014, 7, 3112–3121.
8. S. L. Shen and Q. B. Wang, Chem. Mater., 2013, 25, 1166–1178.
9. M. B. Wilker, K. J. Schnitzenbaumer and G. Dukovic, Isr. J. Chem., 2012, 52, 1002–1015.
10. J. K. Olesiak and H. Weller, ACS Appl. Mater. Interfaces, 2013, 5, 12221–12237.
11. M. D. Regulacio and M. Y. Han, Acc. Chem. Res., 2016, 49, 511–519.
12. T. Torimoto, T. Kameyama and S. Kuwabata, J. Phys. Chem. Lett., 2014, 5, 336–347.
13. B. D. Mao, C. H. Chuang, F. Lu, L. X. Sang, J. J. Zhu and C. Burda, J. Phys. Chem. C, 2013, 117, 648–656.
14. I. Tsuji, H. Kato, H. Kobayashi and A. Kudo, J. Am. Chem. Soc., 2004, 126, 13406–13413.
15. X. N. Liu, Y. H. Tang, S. L. Luo, Y. Wang, X. L. Zhang, Y. Chen and C. B. Liu, J. Photochem. Photobiol., A, 2013, 262, 22–27.
16. B. Liu, X. Li, Q. Zhao, J. Ke, M. Tade and S. Liu, Appl. Catal., B, 2016, 185, 1–10.
17. N. Liang, Q. He, S. Huang, M. Wang, W. Chen, M. Xu, Y. Yuan, J. Zai, N. Fang and X. Qian, CrystEngComm, 2014, 16, 10123–10130.
18. H. Kaga, Y. Tsutsui, A. Nagane, A. Iwase and A. Kudo, J. Mater. Chem. A, 2015, 3, 21815–21823.
19. J. L. Q. Song, T. T. Jiang, T. Y. Guo, L. Liu, H. J. Wang, T. Y. Xia, W. T. Zhang, X. C. Ye, M. Y. Yang, L. X. Zhu, R. X. Xia and X. L. Xu, Inorg. Chem., 2015, 54, 1627–1633.
20. J. Z. Chen, X. J. Wu, L. S. Yin, B. Li, X. Hong, Z. X. Fan, B. Chen, C. Xue and H. Zhang, Angew. Chem., Int. Ed., 2014, 53, 1–6.
21. X. Q. Li, D. L. Wei, S. Z. Kang and J. Mu, Adv. Mater. Res., 2013, 734–737, 2314–2317.
22. T. Kameyama, T. Takahashi, T. Machida, Y. Kamiya, T. Yamamoto, S. Kuwabata and T. Torimoto, J. Phys. Chem. C, 2015, 119, 24740–24749.
23. T. Takahashi, A. Kudo, S. Kuwabata, A. Ishikawa, H. Ishihara, Y. Tsuboi and T. Torimoto, J. Phys. Chem. C, 2013, 117, 2511–2520.
24. Y. Peng, L. Shang, Y. Cao, Q. Wang, Y. Zhao, C. Zhou, T. Bian, L. Z. Wu, C. H. Tung and T. Zhang, Appl. Surf. Sci., 2015, 358, 485–490.
25. Y. J. Yuan, H. W. Lu, Z. T. Yu and Z. G. Zou, ChemSusChem, 2015, 8, 4113–4127.
26. S. Bai, L. M. Wang, X. Y. Chen, J. T. Du and Y. J. Xiong, Nano Res., 2015, 8, 175–183.
27. Y. J. Yuan, Z. T. Yu, Y. H. Li, H. W. Lu, X. Chen, W. G. Tu, Z. G. Ji and Z. G. Zou, Appl. Catal., B, 2016, 181, 16–23.
28. J. F. Xie, H. Zhang, S. Li, R. X. Wang, X. Sun, M. Zhou, J. F. Zhou, X. W. Lou and Y. Xie, Adv. Mater., 2013, 25, 5807–5813.
29. Y. D. Hou, A. B. Laursen, J. S. Zhang, G. G. Zhang, Y. S. Zhou, X. C. Wang, S. Dahl and I. Chorkendorff, Angew. Chem., Int. Ed., 2013, 52, 3621–3625.
30. D. Gopalakrishnan, D. Damien and M. M. Shaijumon, ACS Nano, 2014, 8, 5297–5303.
31. B. D. Mao, Y. B. Yuan, Y. C. Shao, B. Yang, Z. G. Xiao and J. S. Huang, Nanosci. Nanotechnol. Lett., 2014, 6, 685–691.
32. G. H. Tian, Y. J. Chen, Z. Y. Ren, C. G. Tian, K. Pan, W. Zhou, J. Q. Wang and H. G. Fu, Chem.–Asian J., 2014, 5, 1291–1297.
33. Y. Chen, G. Tian, Y. Shi, Y. Xiao and H. Fu, Appl. Catal., B, 2015, 164, 40–47.
34. W. Zhao, Y. Liu, Z. Wei, S. Yang, H. He and C. Sun, Appl. Catal., B, 2016, 185, 242–252.
35. Y. J. Yuan, H. W. Lu, Z. G. Ji, J. S. Zhong, M. Y. Ding, D. Q. Chen, Y. H. Li, W. G. Tu, D. P. Cao, Z. T. Yu and Z. G. Zou, Chem. Eng. J., 2015, 275, 8–16.
36. B. F. Luo, D. B. Xu, D. Li, G. L. Wu, M. M. Wu, W. D. Shi and M. Chen, ACS Appl. Mater. Interfaces, 2015, 7, 17061–17069.
37. X. F. Wang, S. F. Li, Y. Q. Ma, H. G. Yu and J. G. Yu, J. Phys. Chem. C, 2011, 115, 14648–14655.
38. U. A. Joshi, J. R. Darwent, H. P. Yiu and M. J. Rosseinsky, J. Chem. Technol. Biotechnol., 2011, 86, 1018–1023.
39. W. J. Zhang, D. Z. Li, M. Sun, Y. Shao, Z. X. Chen, G. C. Xiao and X. Z. Fu, J. Solid State Chem., 2010, 183, 2466–2474.
40. J. L. Q. Song, T. T. Jiang, G. W. Ji, W. T. Zhang, X. Y. Cheng, W. Weng, L. X. Zhu and X. L. Xu, RSC Adv., 2015, 116, 95943–95952.
41. G. P. Chen, N. Ding, F. Li, Y. Z. Fan, Y. H. Luo, D. M. Li and Q. B. Meng, Appl. Catal., B, 2014, 160–161, 614–620.
42. B. D. Mao, C. H. Chuang, C. McCleese, J. J. Zhu and C. Burda, J. Phys. Chem. C, 2014, 118, 13883–13889.
43. Y. H. Liang, S. L. Lin, L. Liu, J. S. Hu and W. Q. Cui, Appl. Catal., B, 2015, 164, 192–203.
44. X. Xiang, L. Chou and X. Li, Phys. Chem. Chem. Phys., 2013, 15, 19545–19549.
45. A. Boulesbaa, Z. Huang, D. Wu and T. Lian, J. Phys. Chem. C, 2010, 114, 962–969.
46. A. Boulesbaa, A. Issac, D. Stockwell, Z. Huang, J. Huang, J. Guo and T. Lian, J. Am. Chem. Soc., 2007, 129, 15132–15133.
47. K. Li, B. Chai, T. Y. Peng, J. Mao and L. Zan, ACS Catal., 2013, 3, 170–177.
48. K. W. Cheng and S. C. Wang, Mater. Chem. Phys., 2009, 115, 14–20.

---

Footnote

† These authors contributed equally to this work.

---