Improved Mass-Transfer Enhances Photo-Driven Dye Degradation and H2 Evolution over a Few-Layer WS2/ZnO Heterostructure.

In this study, we observed the enhanced photocatalytic activity of a few-layer WS2/ZnO (WZ) heterostructure toward dye degradation and H2 production. The few-layer WS2 acted as a co-catalyst that separated photogenerated electron/hole pairs and provided active sites for reactions, leading to the rate of photocatalytic H2 production of WZ being 35% greater than that over the bare ZnO nanoparticles. Moreover, vortex-stirring accelerated the mass-transfer of the reactants, leading to the efficiency of dye photodegradation being 3 times higher than that obtained without high-speed stirring. We observed a similar effect for H2 production, with greater photocatalytic performance arising from the increased mass-transfer of H2 from the catalyst surface to the atmosphere.

Introduction

In the quest to overcome energy shortages and solve wastewater problems, finding environmentally friendly and low-cost methods remains challenging.[1−8] Among them, semiconductor-based photocatalysts have great potential because they can harvest and convert light into chemical energy.[9−11] For example, transition-metal dichalcogenides are particularly useful, having high intrinsic electrical conductivities, tunable band gaps, and an abundance of active sites for photocatalytic reactions.[12−14] In particular, WS2 is well established as an active photocatalyst for dye degradation and H2 generation. However, the photocatalytic ability of WS2 is highly correlated to its thickness. As the thickness and the size of WS2 decrease, the band gap of WS2 is larger and more active sites are exposed at the edge of the nanosheets, leading to the enhanced photocatalytic efficiency under solar light.[15] Furthermore, WS2 is generally hybridized with other materials for use as a co-catalyst to enhance charge separation and inhibit exciton recombination.[16−19] For instances, Xu et al. demonstrated a 26-fold enhancement of H2 production rate of few-layered 2H-WS2 on CdS, comparing to bare CdS.[20] On the other hand, ZnO is one of the excellent candidate materials for sustainable energy production in consideration of its high photoactivity, low cost, nontoxicity, and stability.[21,22] Nonetheless, its applicability has been limited by fast charge recombination and a wide band gap (ca. 3.37 eV).[23] Therefore, we anchored few-layer WS2 (with sufficient active sites) onto ZnO to facilitate charge separation and photocatalytic reactions, suspecting that we might also improve the ability of ZnO to utilize solar light.[24]

The mixing of reactants can increase their mass-transfer effects and their abilities to interact with substrates, thereby accelerating catalytic reactions.[25−27] Steady mixing can effectively enhance the mass-transfer and mobility of generated reactive species from the catalyst surface (or boundary layers) into solution to form free radicals and, thereby, facilitate catalytic reactions.[28] For example, mass transfer is one of the rate-limiting steps of the H2 evolution reaction (HER).[29] Photocatalytic HER begins with the generation of photoexcited electrons and their transport to the surface of the catalyst. The number of active sites and the charge recombination rate influence the efficiency of proton reduction to produce H2, with mass-transfer of the generated H2 subsequently occurring at the interfaces.[30,31]

Previous studies showed that catalytic performances for wastewater treatment and H2 evolution depend on not only the activity of catalysts but also several parameters such as the mass transfer of reactants and products within the solution.[32,33] However, the effects of the stirring on the photocatalytic efficiency are not yet studied in detail. Herein, we conducted the experiments to understand the stirring effect on photocatalytic abilities for both dye degradation and H2 evolution. The WS2/ZnO (WZ) heterostructures as the photocatalysts were synthesized through the liquid-phase exfoliation (LPE) and solvothermal process. We found that the stirring improves the catalytic efficiency of the WZ heterostructures, and the H2 production rate increases with the increase in stirring speeds and has the optimized value with the appropriate stirring speed.

Results and Discussion

Scanning electron microscopy (SEM) revealed the morphologies of the synthesized catalysts. After LPE (Figure S1, Supporting Information), the few-layer WS2 possessed a nanosheet-like structure (Figure a). We used a solvothermal method to synthesize ZnO nanoparticles (NPs) and annealed them at 250 °C for 4 h (Figure S2); Figure b reveals that they formed aggregated structures with diameters of approximately 50–100 nm. Physically mixing the WS2 nanosheets and ZnO NPs produced the WZ heterostructure. Figure c,d reveals that, after ultrasonication for 30 min, the morphology of the WZ structure featured the few-layered WS2 nanosheets attached to the surfaces of the ZnO NPs. Figure e displays the X-ray diffraction (XRD) patterns of ZnO NPs and the WZ heterostructure. The strong peaks of ZnO matched well with those of the crystal planes of wurtzite ZnO (JCPDS no. 36-1451). For the WZ heterostructure, the major diffraction peaks were those of ZnO, but small peaks appeared for the (002) and (103) planes of exfoliated WS2 (JCPDS no. 08-0237). The relatively low peak intensities of WS2 implied that it was present in low amounts and highly dispersed in the WZ heterostructure. Figure f provides the Raman spectra of the few-layer WS2, the ZnO NPs, and the WZ heterostructure. The characteristic peaks at 351.4 and 417 cm–1 of the few-layer WS2 correspond to its E2g1 and A1g resonance modes, respectively.[34] The peaks at 436.8 and 481.5 cm–1 in the spectrum of the ZnO NPs correspond to its E2 and 2LA modes, respectively.[35] For the WZ heterostructure, the Raman spectrum featured peaks of both ZnO and WS2, revealing the successful synthesis of the heterostructure. To confirm the enhanced light absorption of the WZ heterostructure, we measured the light absorption of the various samples (Figure g). The strong absorption edge of ZnO at 385 nm corresponds to the absorption resulting from its band gap. After ultrasonication with the few-layer WS2 nanosheets to form the heterostructure, the visible-light absorption intensity of the WZ heterostructure was enhanced, consistent with the strong light-harvesting ability of few-layer WS2 in the visible region.[36]

Figure 1

(a–c) SEM images of (a) the few-layer WS2 nanosheets after exfoliation, (b) the ZnO NPs, and (c) the WZ heterostructure. (d) Magnified SEM image of the area in the white box in (c). (e) XRD patterns of the ZnO NPs and WZ heterostructure. (f) Raman spectra and (g) UV–vis spectra of the few-layer WS2, the ZnO NPs, and the WZ heterostructure.

We used transmission electron microscopy (TEM) and atomic force microscopy (AFM) to obtain detailed microstructural information about the WZ heterostructure. Figure a,b presents low-magnification TEM images of the heterostructure, revealing that, after ultrasonication for 30 min, the few-layer WS2 had indeed adsorbed on the surface of the ZnO NPs, consistent with the SEM images (Figure c,d). The high-resolution TEM (HRTEM) image of the WZ heterostructure (Figure c) reveals an exfoliated few-layer WS2 nanosheet having eight layers and a lattice spacing of 0.16 nm, corresponding to the (110) plane of ZnO. Moreover, Figure d and the elemental mapping of WZ (Figure S4) indicate that the units of few-layer WS2 were dispersed well on the ZnO NPs. To prepare the catalysts, we employed LPE, with centrifugation at 7000 rpm, to obtain a few-layered WS2. AFM analysis revealed exfoliated WS2 having lateral dimensions in the range 50–150 nm (Figure e) and a flake having a thickness of 6.695 nm, approximately nine layers (Figure f) if the thickness of monolayer WS2 is 0.7 nm.[37] The thickness distribution of the exfoliated WS2 (Figure g) revealed that, on average, exfoliated WS2 comprised approximately eight layers, with most of the flakes having 3–10 layers, consistent with the results of the TEM (Figure c) and AFM (Figure f) analyses.

Figure 2

(a,d) TEM images of the WZ heterostructure. (b) Magnified TEM image of the area in the white box in (a). (c) HRTEM image of the area in the white box in (b). (e) AFM image of the exfoliated WS2 nanosheets. (f) AFM topographic image revealing that the thickness of exfoliated WS2 was 6.695 nm. (g) Distribution of the number of layers of the exfoliated WS2.

We examined the degradation of MB under a 150-W Hg lamp to evaluate the photocatalytic performance of our various samples. Figure a presents the absorption spectra of MB recorded after irradiation in the presence of the WZ heterostructure; the intensity of the peak centered at 664 nm decreased over time and nearly disappeared after 90 min. Figure b displays the changes in the MB concentration after degradation for various periods of time over the ZnO NPs and the WZ heterostructure; here, C is the remaining concentration of MB at time t and C0 is the initial concentration of MB. Wurtzite-structured ZnO and WS2 have been reported to possess piezoelectrical potentials that can enhance their catalytic abilities.[38,39] We suspected that the improved mass-transfer in the photocatalytic reactions and the piezoelectric potentials produced by the materials were responsible for the superior performance of the stirred catalytic reactions. Accordingly, we examined the degradation in the dark to clarify the influence of light on the catalytic reaction. Because the dye degradation of the samples in the dark was negligible (see the ZnO v and WZ v curves in Figure b), we could eliminate the effect of the piezoelectric potential of the samples in this present system. Figure b reveals that the degradation ratio of the ZnO NPs was only 21.6% after 120 min of irradiation. After loading with the few-layer WS2 to form the WZ heterostructure, the photodegradation ratio increased to 80.5%, confirming the enhancement effect of the few-layer WS2 co-catalyst on the degradation process. The degree of dye degradation was further enhanced when stirring the reaction; the degradation ratios for the catalytic reacts over the ZnO NPs and WZ heterostructure both increased to approximately 63 and 100%, respectively, after 120 min. Figure c presents the temporal degradation curves of MB with respect to −ln(C/C0), fitted using eq

Figure 3

(a) Absorption spectra of MB recorded at various time intervals in the presence of the WZ heterostructure. (b) Degradation efficiency, (c) kinetics, and (d) rate constants of the degradation of MB over the ZnO NPs and WZ heterostructure under irradiation, with and without high-speed stirring (stirring/vibration and irradiation with light are indicated by the descriptors v and l, respectively).

Figure d plots the rate constants of the reactions catalyzed by the ZnO NPs and WZ heterostructure under the various experimental conditions. Here, the rate constant (k) functions as a measure of the catalytic ability. Among our tested samples and conditions, WZ under irradiation and high-speed stirring exhibited the highest catalytic activity, with a value of k (0.04073 min–1) that was approximately 450% greater than that of ZnO (0.00898 min–1) and 300% greater than that of WZ (k = 0.01329 min–1) under irradiation.

We conclude that stirring did not provide a piezocatalytic effect, as mentioned above; nevertheless, it might have enhanced the heterogeneous photocatalytic activity of WZ through improved mass-transfer of reactive oxygen radicals, facilitating their interactions with MB molecules and, thereby, increasing the degradation efficiency. Specifically, the electrons (e–) in the valance band of WZ were excited to the conduction band after absorbing light energy under irradiation; the same amount of holes (h+) was formed in the valence band. The excited electrons reacted with O2 to generate superoxide radicals (O2•–), with hydroxyl radicals produced through h+ being trapped by hydroxyl ions (OH–); these processes are described in eqs –4[40]

Both radicals have great abilities to attract electrons (i.e., great oxidizing ability), thereby degrading the dye molecules. Thus, the removal of MB resulted mainly through radical-mediated reactions. The high velocity of the stirred fluid presumably accelerated the mass-transfer of reactive oxygen radicals from the catalyst surface and boundary layer to the liquid, while also assisting the homogeneous mixing and reaction of the radicals and MB molecules.[28] Sun et al. has demonstrated that a higher rotation speed is beneficial for the uniform flow field distribution and diffusion of MB molecules, resulting in favorable conditions for degradation.[41]

Next, we examined the effect of stirring on the H2 evolution performance, through its effects on mass-transfer and the photocatalytic performance. Figure a,b presents the photocatalytic H2 production obtained over the WZ heterostructure, plotted with respect to time and the stirring speed (from no stirring to 600 rpm), and the corresponding reaction rates. Greater stirring speeds led to greater rates of H2 production. The optimal rate of H2 production was 327 μmol h–1 g–1 when stirring at 300 rpm—approximately 3 times greater than the H2 production of the unstirred counterpart. We attribute this behavior to the increased mass-transfer of generated H2 and the greater probability of contact between the electrons and holes when increasing the stirring speed, both resulting in higher degrees of H2 production.[29,42] Nevertheless, any further increase in the stirring speed did not benefit the reaction, presumably because unstable rotation of the magnetic stirrer impeded mass-transfer. Indeed, the visibly turbulent flow of the solution at 600 rpm decreased the rate of H2 production to slightly less than that observed at 100 rpm. Figure S5 compares the rates of H2 production over the bare ZnO NPs and the WZ heterostructure at 300 rpm. The rate of H2 production was improved 35% merely by loading a low amount of the few-layer WS2 (0.68 wt %) on the ZnO NPs. Figure c displays the excellent repeated ability of WZ heterostructures. It is clear that WZ retains photocatalytic ability without remarkable decline after three repeated tests. Table shows the comparisons of the photocatalytic H2 production of the WZ heterostructure with the ZnO-based study in the literature. Obviously, the H2 production in the present study is comparable with that of ZnO-based photocatalysts in the literature.

Figure 4

(a) Photocatalytic H2 evolution and (b) rates of H2 evolution over the WZ heterostructure at various stirring speeds. (c) Repeated tests of the regeneration of WZ heterostructures for photocatalytic H2 evolution at 300 rpm. (d) Schematic representation of H2 evolution over the WZ heterostructure with Na2S and Na2SO3 as sacrificial agents.

Table 1

Comparison of Photocatalytic H2 Evolution of WZ Heterostructure and Other ZnO-Based HER Photocatalysts

photocatalyst	hydrogen production rate (μmol h–1 g–1)	sacrificial reagents	light source	refs
WS2/ZnO	327	Na2S/Na2SO3	300-W Xe lamp (with AM-1.5)	this work
ZnO/ZnS heterostructure	415.3	Na2S/Na2SO3	300-W Xe lamp (with λ ≥ 420 nm)	(43)
ZnO/ZnS heterostructure	384	glycerol	350-W Xe lamp	(44)
g-C3N4/ZnO	322	triethanolamine	300-W Xe lamp (with λ ≥ 420 nm)	(45)
ZnO/ZnxCd1–xTe	265	Na2S/Na2SO3	300-W Xe lamp (250 < λ < 780 nm)	(46)
3.0 wt % MoS2/ZnO	165	Na2S/Na2SO3	300-W Xe lamp (with λ ≥ 400 nm)	(47)

Figure d presents a possible mechanism for the photocatalytic reaction for H2 production over the WZ heterostructure. The experimental results can be ascribed to two factors: (i) better charge separation and more active sites for the HER provided by the few-layer WS2 co-catalyst on the ZnO NPs; (ii) accelerated mass transfer resulting from higher rates of stirring. The band alignment of WZ facilitates the transfer of carriers from the conduction band (CB) of ZnO to the CB of WS2, thereby prolonging the lifetime of the photoexcited carriers. Once the carriers are produced in the materials through the absorption of light, the electrons transfer to the CB of the few-layer WS2. The exfoliated WS2 not only served as a co-catalyst to enhance charge separation but also exposed an abundance of active edge sites for the HER.[36] On the other hand, stirring also enhanced the photocatalytic H2 production through two possible mechanisms. First, mass-transfer in the suspension will be accelerated upon increasing the speed of stirring; a higher rate of mass-transfer of generated H2 from the catalyst surface to the atmosphere would benefit H2 production and diminish the backward reaction of H2.[29] Second, a greater speed of stirring will accelerate the flow rate of the solution, thereby increasing the probability of contact between the photoexcited electrons and protons. The increased accessibility of H+ ions within the solution for reaction with electrons on the catalyst surface might also enhance the HER performance. In general, the approach of speeding up the magnetic stirring is an effective and simple means of further boosting the efficiency of the photocatalytic HER.

Conclusions

In summary, we have fabricated ZnO NPs and few-layer WS2 through a solvothermal method and LPE, respectively, and then used ultrasonication to obtain a WZ heterostructure. With the addition of the few-layer WS2 as a co-catalyst, the heterostructure exhibited enhanced photocatalytic ability, which we attribute to WZ displaying enhanced charge separation, an abundance of active sites, and improved absorption in the visible region. A study of the mass-transfer effect revealed improved photocatalytic dye degradation (k = 0.04073 min–1) and the highest rate of H2 generation (327 μmol h–1 g–1) being obtained. Accordingly, we conclude that the photocatalytic performance was boosted by accelerated mass-transfer of the reactants and products, as well as the increased probability of contact between the reactants. We hope that the insights obtained from this study may initiate new practical ways to perform wastewater treatment and the photocatalytic HER.

Experimental Section

Chemicals

Tungsten disulfide (WS2, powder, 2 μm, 99%, Sigma-Aldrich), zinc nitrate hexahydrate [Zn(NO3)2·6H2O, 99%, Alfa Aesar], potassium hydroxide (KOH, 85%, ECHO), ethylenediamine (C2H8N2, Sigma-Aldrich), and ethanol (C2H5OH, 99.5+%, ECHO) were used without further purification.

Liquid-Phase Exfoliation of WS2

LPE was used to obtain a few-layer WS2 (Figure S1). Bulk WS2 (400 mg) was added into 35% ethanol (50 mL) as a dispersive agent and then the mixture was ultrasonicated for 16 h at a power of 100 W and a frequency of 37 kHz. The suspension was then centrifuged (7000 rpm, 10 min) twice to separate the exfoliated structures, and the supernatant of yellowish-brown color was collected to give a few-layer WS2 solution. The concentration of the WS2 nanosheet solution was kept at 6.785 ppm for the formation of the heterostructures, and the concentration was determined by ICP–MS (Agilent 7500ce).

ZnO Nanoparticles

ZnO nanoparticles (NPs) were synthesized using a solvothermal method. Zn(NO3)2·6H2O (1.19 g, 4.00 mmol) and KOH (0.226 g, 4.00 mmol) were dissolved in 95% ethanol (20 mL) under continuous stirring. Ethylenediamine (0.3 mL) was added and then the mixture was stirred for 10 min. The as-prepared solution was transferred into a Teflon-lined stainless-steel autoclave. The autoclave was heated to 120 °C at a rate of 5 °C//mmin and then kept at 120 °C for 16 h. After cooling, the suspension was washed several times with ethanol and deionized water to collect the precipitate, which was dried at 60 °C for 6 h. The as-synthesized ZnO was then annealed at 250 °C for 4 h.

WS2/ZnO Heterostructure

The few-layer WS2 solution (50 mL) and the ZnO NPs (50 mg) (1:1) were mixed in a beaker (Figure S2). The mixture was ultrasonicated for 30 min to induce physical adsorption. The color of the solution changed from translucently brown to opaquely brown when adding the ZnO NPs into the WS2 nanosheet solution, and the solution became transparent after stored in the dark for 1 day, which implies the heterostructures formed and then precipitated at the bottom of the solution (Figure S3). WS2/ZnO-heterostructured sediment was collected and dried at 60 °C for 6 h.

Characterization

The crystallographic properties of the catalysts were characterized using XRD (Bruker D2 phaser) with Cu Kα radiation. The morphologies were observed using SEM (Hitachi SU8010); detailed information was obtained using HRTEM (JEOL ARM-200F) and a Cs-corrector. The thickness of the samples was measured using AFM (Bruker Dimension Icon). Raman spectra were recorded at room temperature using a Micro-Raman spectrometer (ProTrusTech, MRI-1532A) and a 532-nm excitation laser. UV–vis absorption spectra of the samples were recorded using a UV–vis spectrometer (Avantes AvaSpec-2048UV spectrum).

Photocatalytic Degradation of MB

The photocatalysis test of WS2/ZnO (WZ) was performed by using methylene blue (MB, C16H18ClN3S, also known as methylthioninium chloride). The catalyst (2 mg) was dispersed in a MB solution (5 ppm, 5 mg/L, 15 mL). The mixture was stirred magnetically for 20 min at room temperature in the dark to establish an adsorption/desorption equilibrium between the photocatalyst and MB. Degradation tests were performed under irradiation with light from a 150-W Hg lamp and continuous stirring. A vortex mixer (MX-S) was used to provide extra high-speed stirring. The dye degradation measurements were carried out with the pH value of around 5–6 at room temperature. At different intervals during the degradation process, samples (1.5 mL) of the suspension were collected and centrifuged (10,000 rpm, 5 min) to settle the catalyst at the bottom of the tube. The absorption spectrum of the supernatant and the concentration of the remaining MB were analyzed using a UV–vis spectrometer (Avantes AvaSpec-2048 UV spectrum).

Photocatalytic H2 Evolution

A photocatalytic H2 evolution was tested in a glass vial (20 mL) closed with a sleeve stopper. The catalyst (3 mg) was dissolved in a mixture of Na2S and Na2SO3 (0.45 M, 10 mL) in the vial with magnetic stirring. The vial was purged with Ar for 15 min to eliminate air. The H2 evolution test was performed under a 300-W Xe lamp with an AM-1.5 filter and continuous stirring. The speed of magnetic stirring was varied to test the relationship between the dispersity and the photocatalytic activity. The total H2 produced in the reactor was measured using gas chromatography (GC, Shimadzu GC-2014).