Semiconductor metal oxides photocatalysts, such as TiO2,(1; 2) ZnO,(3) and SnO2,(4) have been extensively studied because of their non-toxicity, chemical inertness, photostability, and ability to utilize visible and UV light. The main disadvantages that hinder the practical application of single phase semiconductor photocatalysts are the fast recombination of photogenerated electron–hole (e-/h+) pairs and the narrow light-response region due to their relatively large band gaps (e.g., 3.37 eV for ZnO).(5; 6) Transition metal dichalcogenides (TMDs), with direct band gaps ranging from 1 to 2.5 eV, have emerged as promising materials for photocatalysis in the near-infrared to visible range.(7-9) For example, molybdenum disulfide (MoS2), one of the most stable layered TMDs, is considered an emerging material for efficient photocatalysis.(10) MoS2 nanosheets have a direct band gap of 1.9 eV, appropriate for activating photocatalytic processes by visible light. Further advantages include high in-plane electron mobility, nontoxicity, and earth-abundance. Nevertheless, TMDs also suffer from rapid photoelectron-hole recombination.(11) A strategy for increasing photoelectron and hole lifetimes is to use quantum confinement effects by reducing the dimensionality of metal oxide structures and creating core-shell or layered heterostructures. For example, core-shell TiO2-ZnO spherical particles demonstrated reduced electron-hole recombination and a red shift of the absorption band to the visible region, leading to an enhanced photoconversion efficiency of 0.55%.(12) Computational predictions show even higher potential photocatalytic activity in 2D layered heterostructures with type II band alignment.(13) In particular, a photocatalytic efficiency of 1% is predicted in MoS2-ZnO and WS2-ZnO layered heterostructures, in which MoS2 (or WS2) predominantly contributes to the conduction band minimum (CBM) and the ZnO layers dominate the valence band maximum (VBM).(13) This physical separation of photoelectrons and holes in different layers of a heterostructure significantly reduces charge recombination. Furthermore, the reduced bandgaps of these heterostructures, 1.60 eV for MoS2-ZnO and 2.05 eV for WS2-ZnO, promote more efficient visible light absorption than bulk ZnO.
To exploit the emergent electronic and photocatalytic properties of layered heterostructures in various applications, reliable and scalable synthesis approaches must be developed.(14) Existing methods suffer from low reproducibility and are labor intensive. For example, the most common synthesis approach to produce van der Waals heterostructures from naturally layered materials with relatively weak van der Waals interactions between the layers involves exfoliation and manual stacking of layers of unlike materials. This method was widely used in proof-of-concept experiments, but cannot be scaled up due to poor control over the mutual orientation of the layers(15) This approach is also not applicable to the synthesis of heterostructures where at least one component has ionic or covalent bonding in all three crystallographic directions. Synthesis of 2D layers of materials that are three-dimensional (3D) at standard conditions can be divided into two categories: 1) chemical vapor deposition on metal substrates and 2) hydrothermal synthesis in the presence of ligands. Chemical vapor deposition furnishes high level of control over the quality and structure of layered materials.(15) A notable application of this method was the synthesis of graphene-like ZnO (g-ZnO) on various metal substrates.(16) The limitations of chemical vapor deposition, however, include difficulties in lifting the 2D layers off the substrate and poor scalability. In contrast, hydrothermal synthesis is highly scalable. 2D ZnO produced under mild hydrothermal conditions in aqueous solutions showed the presence of single layers and bi-layers of ZnO in the samples, observed by high resolution transmission electron microscopy (HRTEM) imaging.(17; 18) Thus, the hydrothermal method furnishes an attractive alternative approach for synthesizing multilayer semiconductors for functional devices. Previous studies also demonstrated the feasibility of hydrothermal synthesis of MoS2/ZnO heterostructures.(19) However, the reported materials have polycrystalline structure with crystallites composed of MoS2 nanoflakes decorated with ZnO nanoflakes and agglomerated into sub-micron size microsphere. Alternative approaches to produced layered MoS2/ZnO heterostructures used hybrid organic-inorganic layered ZnO (LHZnO), which was mechanically mixed with single layer MoS2,(20) co-precipitation method,(21) and hydrothermal synthesis, which used MoS2 flakes and ZnO nanorods as precursors, and produced composites with MoS2 flakes connected to ZnO nanorods through the ZnS interlayer.(22) Similar templating approach was used to produce MoS2 decorated TiO2 nanobelts (23) and C-ZnO/MoS2 nanocomposite anchored on 3D mesoporous carbon framework (24) All four methods produced ZnO-MoS2 heterostructures with enhanced visible-light photocatalytic activity, however the level of control over materials morphology, crystallinity, and mutual orientation between MoS2 and ZnO layers was limited.
This study proposes a novel solution-based approach, in which 2D dichalcogenide layers template the growth of metal oxides to promote the development of 2D layered architectures through intermediate phase.(25) By doing so, we can achieve the envisioned 2D architecture to overcome the limitations of three-dimensional semiconductors, which involves the transformation of bulk metal oxides into graphene-like sheets, effectively reducing their intrinsic dimensionality. The resulting quantum confinement is expected to produce heterostructures with type II band alignment and enhance the lifetime of photogenerated electron and holes. The choice of ZnO and MoS2 for the proof-of-concept synthesis was determined based on the known photocatalytic properties of each individual component and the predicted improved performance of the heterostructures. Specifically, graphene-like or multilayer ZnO heterostructures with transition metal dichalcogenides are predicted to have an enhanced photocatalytic activity in the visible to UV range.(26) To address this major synthetic challenge, we utilize 2D templated ZnO growth with MoS2 nanosheets produced via a liquid phase ultrasonic exfoliation method. The ZnO-MoS2 hybrid structure is then synthesized by stabilizing the intermediate phase of Zn(OH)2 plates on MoS2 flakes followed by thermal annealing. The hybrid structure exhibits enhanced photocatalytic activity toward the degradation of an organic dye. The superior performance over pristine MoS2 and bare ZnO nanoparticles can be attributed to the enhanced separation efficiency of photogenerated charge carriers. The facile, low-cost, and low-temperature solution-based method reported here exemplifies a strategy of hybridizing nanoscale 2D exfoliated defect-free nanosheets with a wide selection of semiconductor nanostructures and furnishes a new paradigm for the synthesis of functional materials for clean energy applications.
Figure 1. SEM images of Zn(OH)2/ZnO obtained by the hydrothermal method at 60°C for 24 hours in the presence of different MoS2 content: (A) 1% and (B) 5% when controlling the molar ratio of Zn(NO3)2●6H2O to HMTA at 200: 20 and the molar ratio of water to IPA at 5:1; (C) Powder XRD pattern of the exfoliated MoS2 and final zinc products synthesized under different controlled conditions; (D) A scheme showing the synthesis procedure for MoS2-Zn(OH)2/ZnO.
In this study, ZnO was synthesized in a pH buffer using zinc nitrate as the Zn2+ precursor and HMTA as a weak base. During the reaction, HMTA slowly decomposes in the heated aqueous solution to yield ammonia and formaldehyde as the initial reactants. These in turn result in the formation of hydroxide ions (Eq. 1-2):
The introduced OH-groups can bind to the MoS2 surface (2H phase), forming a S-Mo-S-OH layer. The MoS2-OH layer then acts as a reactive template for heterogeneous nucleation of zinc hydroxides through the reaction of hydroxide ions with Zn2+(Eq. 3). The morphology of the prepared binary composite MoS2/Zn(OH)2/ZnO at different MoS2 content (Figure 1A,B and Figure S1) changes from irregular thin plates to hexagonal plates with larger surface area and thickness was MoS2 content increases from 1% to 5%. At 1% MoS2, a mixture of large and small plates with thicknesses from 200-300 nm was observed (Figure 1A). When MoS2 content was increased to 5%, regular larger plates with a thickness of ~500 nm was observed (Figure 1B), highlighting the role of the bilayer MoS2-OH in directing the formation of large surface area Zn(OH)2/ZnO plates. XRD analysis was conducted to identify the composition of the samples, shown in Figure 1C. The (102) peak of MoS2 (dashed line) is observed in the XRD pattern of the composite and slightly downshifts as the molar ratio of MoS2 in the composites increases, indicating binding between MoS2 and Zn(OH)2/ZnO. The conclusion is further supported by the results of control experiments, in which MoS2 was absent from the system (Figure S2A), the HMTA was replaced with NaOH (Figure S2B), or the mixed IPA-aqueous solvent was replaced with pure water (Figure S2C). In all these control experiments only pure ZnO twin structures or nanoparticles were obtained, as evidenced by the characteristic XRD pattern (Figure 1C, green, purple and yellow curves). Thereby, the synergistic effect of Zn(NO3)2 and HMTA in a 200:20 ratio, IPA-water mixed solvents, and MoS2 is crucial for the formation of a 2D heterostructure (see SI for details, Figures S3-S6). Furthermore, in the presence of MoS2, increasing of concentration of HMTA from 50mM to 200mM resulted in the formation of zinc hydroxonitrate Zn5(OH)8(NO3)2(H2O)2 in addition to Zn(OH)2 and ZnO phases. The observed formation of phases with higher OH content is likely due to the enhanced release of OH groups produced during HMTA decomposition and enhanced functionalization of MoS2. The hydroxylated S-Mo-S-OH template likely stabilizes intermediate phases and inhibits their conversion into ZnO. Previous studies of ZnO growth in the presence of HMTA revealed that HMTA has a pH buffering effect and that the maximum rate of growth was achieved when HMTA supersaturation slightly exceeded that of Zn(NO3)2.(27) Our experiments showed a similar trend. Further increasing the HMTA concentration to 500 mM at a fixed Zn(NO3)2 content resulted in a higher fraction of ZnO than the Zn5(OH)8(NO3)2(H2O)2 phase (Figures S3 and S4). On the other hand, decreasing Zn(NO3)2 concentration from 200 mM to 20 mM at constant HMTA content in the absence of MoS2 led to the formation of pure ZnO twinned crystals (Figures S5 and S6). Under the same conditions in the presence of a MoS2 templating surface, the fraction of ZnO decreased and that of Zn(OH)2 increased (Figures S5 and S6), supporting the proposed mechanism of the nucleation of an intermediate 2D ZnO phase on a S-Mo-S-OH template.
The final stable ZnO-MoS2 composite phase was obtained by thermal treatment of the Zn(OH)2 /ZnO /MoS2 at 350 °C. Annealing did not alter the original layered structure of the composites observed in SEM micrographs (Figure 2,A1-D1). No intermediate phase was seen in the XRD patterns for all of the annealed products (Figure 2F), confirming the thermal conversion of Zn5(OH)10·2H2O to pure ZnO. High-resolution SEM images (Figure 2,A2-D2) indicate that the ZnO plates are not homogeneous and are composed of different sized grains sintered together to form a relatively compact plate. The surface grain size, calculated from the SEM images (Figure 2,A3-D3), is smaller for the ZnO-MoS2 heterostructure than samples synthesized without MoS2. The grain size difference further illustrates the templating effect of MoS2 on the nucleation and growth of ZnO.
Figure 2. SEM image of porous ZnO-MoS2 heterostructure after annealing. (A, B, C) Zn(NO3)2:HMTA=200:20 with 5% MoS2; (D, E, F) Zn(NO3)2:HMTA=50:20 with 5% MoS2; (G, H, I) Zn(NO3)2:HMTA=200:100 with 5% MoS2; (D1, D2) Zn(NO3)2:HMTA=200:200 with 5% MoS2; (J, K, L) Particle size distribution for each heterostructure derived by the corresponding conditions. (M) EDS mapping distribution for the image in C1; (N) XRD pattern of annealed samples synthesized at different conditions.
The structure of the obtained ZnO-MoS2 composite was analyzed using HRTEM. The imaging revealed large surface area (0001) ZnO plates on the surface of the heterostructure (Figure 3A, inset). The examination of the atomic structure of the surface layers revealed distinct regions with varying interplanar spacing. Specifically, interplanar spacings of 0.27 nm and 0.16 nm were identified, corresponding to the(100) and (110) facet of MoS2, respectively. Additionally, regions with an interplanar distance of 0.26 nm were observed, attributed to the (001) facet of ZnO.. The structure of the ZnO-MoS2 plates was further analyzed using cross-sectional S/TEM imaging (Figure 3C,E), revealing the presence of periodically alternating layers. The micrograph shows a layered structure of interchanging regions with characteristic ZnO (110) and MoS2 (110) spacings where each region is composed of 66 ± 8 and 9 ± 2 layers, respectively. As shown in Figure 3D and F, the ZnO (110) 2D regions are not perfectly parallel to each other, but form a relative angle varying between 0.9º and 4.5º. EDS mapping shows that areas containing higher ZnO density have less MoS2 (yellow rectangles in (G)) and that MoS2 is absent in regions with no ZnO (orange rectangles in (G)). Combined, these data confirm the formation of a layered heterostructure with interchanging 2D regions of ZnO and MoS2. The alignment of the ZnO (110) and MoS2 (110) facets in the cross-section of heterostructure suggest an in-plane alignment of ZnO and MoS2 along the (001) direction. Although the TEM data are insufficient to make definitive conclusions about the details of the lateral alignment between the ZnO and MoS2 layers, the observed in-plane orientation is consistent with epitaxy between the components of the heterostructure. Specifically, the lattice parameters of ZnO and MoS2 in the a- and b- directions are almost identical and equal to a = b = 3.24 Å for ZnO and a = b = 3.19 Å for MoS2, respectively. This 1.5 % lattice mismatch supports the formation of epitaxial heterostructures along the (001) direction, however, further studies are needed to confirm the epitaxial matching between the layers.
Based on these results, two possible pathways to heterostructure formation can be proposed. One possibility is the direct nucleation of ZnO or Zn(OH)2 on functionalized MoS2-OH flakes followed by the assembly of ZnO/MoS2 flakes into a layered structure. Alternatively, ZnO can epitaxially grow in the confined space between MoS2 flakes. In both scenarios HMTA and IPA serve as pH regulating agents and sources of OH-groups(27) for the formation of ZnO and functionalization of the MoS2 surface. Current ex situ studies do not provide sufficient evidence to delineate between these pathways and future studies will focus on elucidating the mechanism of ZnO nucleation on MoS2.
Figure 3. TEM and HRTEM images of a ZnO-MoS2 heterostructure (A, B); TEM image of the cross-sectional view of the heterostructure (C); HADDF-STEM images of ZnO-MoS2 (D)(E)(F); EDS mapping of ZnO-MoS2 heterostructure(G).
The photocatalytic activity of the layered ZnO-MoS2 heterostructure was investigated using the photocatalytic degradation of Rhodamine B (RB) as a model reaction. The catalyst and dye solutions were aged in the dark for 30 min to establish the adsorption/desorption equilibrium between the dye and ZnO-MoS2 catalyst prior to irradiation with visible light. The adsorption spectra of RB were recorded during the 90 min irradiation of the sample under a 150W Xe-lamp (Figure S8 A-D). The intensity of the main absorption peak located at 550 nm continuously decreased as irradiation time increased. The RB solution gradually became colorless during irradiation, indicating the gradual degradation of the dye. The rate of photodegradation was largest for ZnO-MoS2 heterostructure relative to other photocatalysts. The rate of degradation follows the Langmuir-Hinshelwood pseudo-first-order model:
-ln(C/C0) = K·t + c
Where K is the pseudo-first-order degradation rate constant, and t is the specific time. The kinetics were measured by plotting the -ln(C/C0) versus t. C is a constant.
The observed pseudo-first-order degradation rate constants (K) for four samples(S1, S2, S3 and S4) were as follows: S1 (Zn(OH)2 mixed with MoS2) - 0.0005 min⁻¹, S2 (ZnO mixed with MoS2) - 0.019 min⁻¹, S3 (Zn(OH)2-ZnO-MoS2) - 0.012 min⁻¹, and S4 (ZnO-MoS2 heterostructure) - 0.023 min⁻¹ (Figure 4A). A comparative analysis of the photodegradation extent of RB in the presence of different catalysts, including pure Zn(OH)2, pure ZnO, Zn(OH)2 mixed with MoS2, ZnO mixed with MoS2, Zn(OH)2-ZnO-MoS2, and the ZnO-MoS2 heterostructure, revealed a remarkable nearly 50% improvement in the photocatalytic efficiency of the layered heterostructure compared to that of the pure components or their mixtures (Figure 4B).
Photocatalysis is a multifaceted process influenced by various material properties, including surface area, crystallinity, and the electronic structure of catalysts.(28; 29) The role of surface area and crystallinity is undoubtedly crucial when assessing photocatalytic performance. Nevertheless, in our investigation, all materials (S1, S2, S3, S4) exhibited a consistent lack of significant porosity, indicating a uniform surface area profile across the samples. Additionally, the materials showcased outstanding crystalline properties(Figure 1C), thereby minimizing variations related to crystallinity. This uniformity allows us to isolate and underscore the impact of electronic states on the enhanced performance of sample 4. In the case of MoS2 and ZnO, the efficient electron transfer process facilitated by the amalgamation of MoS2 and ZnO, resulting in a greater quantum efficiency for generating reactive species, such as reactive oxygen species (ROS). These ROS are pivotal intermediates responsible for pollutant degradation and other desired reactions. Furthermore, it is pertinent to note that the interplay of electronic states in the hybrid material may also confer stability and durability, mitigating issues like photocorrosion and ensuring the long-term sustainability of the enhanced performance. In a word, while we acknowledge the significance of surface area and crystallinity in the context of photocatalysis, our extensive characterization and analysis compellingly affirm the central role of the electronic states of MoS2 and ZnO within the hybrid material. The synergistic interplay between their electronic structures underlines the marked improvement in photocatalytic performance(sample 4).
The observed significant improvement in the photocatalytic efficiency of the epitaxial ZnO-MoS2 heterostructure is also in line with theoretical predictions for layered ZnO-MoS2. Density functional theory simulations predict that ZnO interfacing MoS2 retains its wurtzite structure if the ZnO is at least 4 monolayers thick.(30) These epitaxial ZnO-MoS2 interfaces are predicted to have a type II band alignment with the CBM associated with Mo 4d states of MoS2, a VBM mostly with O 2p states of ZnO with a small contribution from Mo 4d, and a bandgap of 1.33 eV (Figure 4F). (13; 30; 31) This band alignment promotes the physical separation of photogenerated electrons and holes in MoS2 and ZnO phases, respectively, thereby reducing the probability of charge recombination. The second factor that could affect the photocatalytic activity of the heterostructure is the higher specific surface area of the photocatalyst compared to the corresponding pure phases or previously reported core-shell structures.(32) As seen from Figures 3E and 3B, the relatively thin and small surface area MoS2 nanosheets form an epitaxial heterostructure of relatively high surface area with the extended ZnO nanosheets. The extended layered structure enables a higher adsorption of the dye molecules on the heterostructure surface with a resulting enhancement in photocatalytic response. Together, the enhanced specific surface area and lower photoelectron/hole recombination of the catalyst promote more efficient generation of the active species, such as O2-, •OH , e-, and h+, for dye decomposition. (29) These data suggest that the robust interaction between ZnO and MoS2 sheets within the nanocomposite fosters effective charge separation and expedites charge transfer processes.
Figure 4. Photocatalytic activity of ZnO and MoS2 catalysts. Photoluminescence spectra of different catalysts: (A) S1: Zn(OH)2 mixed with MoS2; S2: ZnO mixed with MoS2; S3: Zn(OH)2-ZnO-MoS2; S4: ZnO-MoS2 heterostructure. (B) Effects of different catalysts on the photocatalytic degradation efficiency of RB under 150 W Xe-lamp. (C) The photocatalytic reaction pathway for dye decomposition and predicted band alignment in the layered ZnO-MoS2 heterostructure.
In this work, a scalable hydrothermal method was developed for the synthesis of a 2D layered ZnO-MoS2 heterostructure through stabilizing intermediate states of Zn(OH)2 on S-Mo-S-OH. Characterization of the structure and morphology indicates that the MoS2 flakes are intercalated between large size ZnO plates to form an epitaxial layered heterostructure. The resulting quantum confinement and band alignment was shown to improve the photocatalytic efficiency of the layered material by almost 50%, compared to the corresponding single-phase catalysts and mixtures. The proposed synthetic approach furnishes a new general method for the scalable and reproducible synthesis of 2D layered heterostructures through stabilizing intermediate states.