Preparation and application of 0D, 2D and 3D molybdenite: a review

Photoluminescence

Compared with the quenching properties of MoS₂^[8], one of the most promising properties of 0D MoS₂ is photoluminescence (PL). The PL emission spectra of 0D MoS₂ are at different excitation wavelengths (320-520 nm) compared with that of bulk MoS₂^[9]. With a change in excitation wavelength, the emission wavelength also changes, exhibiting an excitation wavelength-dependent effect. Fluorescence microscopy images of dried 0D MoS₂ at 405, 488 and 552 nm show that 0D MoS₂ emits blue, green and red light, respectively.

In addition, there exist some relationships between the PL intensity of 0D MoS₂ and pH. For example, 0D MoS₂ synthesized by Wang et al.^[10] showed that its PL is almost unaffected by pH. However, the PL of 0D MoS₂ synthesized by Dai et al.^[11] was basically influenced by pH. At pH 2-6, the PL of 0D MoS₂ increases with increasing pH. However, the PL decreases with increasing pH without the movement of the peak at pH 10-12. Interestingly, Siddiqui et al.^[12] prepared 0D MoS₂ with excitation wavelength independence. When the excitation wavelength varied from 260 to 350 nm, the emission wavelength was 448 nm.

Anti-Stokes luminescence

Anti-Stokes luminescence is a short wavelength and high frequency light excited from long wavelength, low frequency excitation light. Dong et al.^[13] reported the anti-Stokes luminescence properties of 0D MoS₂ in the wavelength range of 775-900 nm. When the excitation wavelength changed from 775 to 900 nm, the corresponding emission peak moved from 525 to 610 nm.

Cytotoxicity

Since 0D MoS₂ can be easily dispersed in water and exhibits strong fluorescence properties, concerns have been raised regarding its cytotoxicity. For instance, Xu et al.^[14] used the MTT method to detect the cytotoxicity of 0D MoS₂ and HEK 293T and HeLa cells with different concentrations of 0D MoS₂ were incubated for 24 h to conduct cell survival experiments. The results showed that the survival rate of the HEK 293T and HeLa cells decreased by ~1% and ~11%, respectively, when the concentration of 0D MoS₂ was 250 µg/mL. When the concentration of 0D MoS₂ was increased to 2000 µg/mL, the cell survival rate of HEK 293T cells decreased by ~12% and that of HeLa cells by ~30%, revealing that 0D MoS₂ was harmless to the cells. The results showed better biocompatibility compared to the cytotoxicity of MoS₂ reported by Wang et al.^[15] and also indicated the potential application value of 0D MoS₂ in the field of bioimaging.

Electrochemical activity

Recently, researchers have attempted to explore the high catalytic behavior of 0D MoS₂ in the hydrogen evolution reaction. Qiao et al.^[16] synthesized 0D MoS₂ via multiple lithium intercalation and stripping and then tested its electrocatalytic activity using 0.5 M H₂SO₄ as an electrolyte in a typical three-electrode system. The results showed that 0D MoS₂ prepared with more stripping cycles had a lower overpotential, higher cathode current and lower Tafel slope. The reason for this was the existence of the rich active edge of 0D MoS₂ that endows it with excellent catalytic performance, thereby representing a new method for the application of 0D MoS₂ in electrochemistry.

Electronic structure

MoS₂ is a unique layered transition metal sulfide and its band gap varies with the number of layers. The direct band gap of bulk MoS₂ is ~1.29 eV^[18]. Due to its ultrathin 2D crystal structure, the band structure and band gap of MoS₂ are affected by the quantum local effect^[19]. The highest valence band (VB) is at the Г point and the lowest conduction band (CB) is about half of the Г-K direction, which causes the transition from an indirect to direct band gap^[20]. As the number of layers decreases, the lowest band in the CB rises, thereby increasing the overall bandwidth. The K point in the Brillouin region is mainly due to the fact that the d orbital of the Mo atom is not affected by the interlayer interaction. The direct band gap only increases by ~0.05-0.10 eV at the K point in the Brillouin region^[21]. The CB state near the Г point is caused by the interlayer interaction between the P_z orbital of the S atom and the d orbital of the Mo atom. Therefore, the band gap is more susceptible to the layer number change at the Г point^[22]. When the number of layers becomes one (i.e., a single layer), the indirect conversion band gap is larger than the direct conversion band gap and the minimum direct band gap is ~1.9 eV at the K point. This special transition from indirect to direct band gap originates from the interlayer influence and quantum local and long-range Coulombic effects^[23].

Optical properties

Photoexcited optical studies show that monolayer MoS₂ has a very high quantum yield with a photoconversion band gap of ~1.9 eV. Compared with bulk MoS₂, the quantum yield of 2D MoS₂ is very high at 4 × 10^-3[20]. Thus, photoemission and Raman spectroscopy can be used to characterize the quality and crystallinity of 2D MoS₂^[24]. The photoemission spectrum of 2D MoS₂ has two exciton peaks at the K point, namely, the exciton A peak and exciton B peak, which may be caused by the spin orbit splitting of VB. VB splitting occurs at the K point, the highest point in the VB. Therefore, two conversions from excitons to the double degenerate CB are possible^[25]. 2D MoS₂ possesses higher mechanical strength and Young’s modulus than steel^[26]. The distortion of 2D MoS₂ can reach 11% without rupture^[27]. In addition, 2D MoS₂ exhibits piezoelectricity and can be used in sensitive sensors^[28]. Theoretical calculations have shown that the effective electron mass of 2D MoS₂ at the K point is 0.48 m_e, much larger than the value of 0.012 m_e for graphene^[29]. Feng et al.^[30] created a broadband optical funnel by adjusting the tension of 2D MoS₂.

Characterization

Researchers have designed and constructed a series of 3D MoS₂-based materials with various frameworks (e.g., aerogels, sponges and foams). These 3D MoS₂-based materials not only retain the intrinsic properties of MoS₂ but also possess some new collective physiochemical properties, including high porosity, low density, large specific surface area, strong mechanical strength and unique electrochemical performance. Furthermore, 3D MoS₂-based materials obtain multi-level pore structures (micropores, mesopores and macropores) and internal connection frameworks, which endow them with improved surface accessibility and better mass transport.

Zhang et al.^[31] prepared a MoS₂/graphene foam using nickel foam as a template. Nickel foam itself is a porous, microporous structure and the MoS₂/graphene obtained by this method has a large specific surface area with a Tafel slope of only 42.8 mV/decade. The existence of the 3D structure not only promotes the electron transfer process but also effectively avoids the volume change of the electrode for long periods in electrocatalytic hydrogen evolution, thereby enhancing the stability of the electrode. Liu et al.^[32] synthesized a MoS₂/graphene hydrogel via a one-step hydrothermal method, which confirmed the existence of a microporous structure by scanning electron microscope (SEM) and specific surface area test, and directly used the prepared material as the electrode for an electrocatalytic hydrogen evolution reaction. The electrochemical experiments showed that MoS₂ has a lower initial potential and Tafel slope and its current density reached 52.4 A/g at a potential of 400 mV.

Liquid exfoliation method

The liquid exfoliation method uses ultrasonication to disperse bulk MoS₂ in an organic solvent and ultrasonically treat the suspension. During the ultrasonication process, MoS₂ is damaged and sheared into small sizes by mechanical force^[34]. The choice of solvent is particularly important because the solubility and surface tension index of the solvent are related to the good monodispersity^[35]. Gopalakrishnan et al.^[36] subjected a MoS₂ powder to bath sonication and probe sonication in sequence for 3.5 h in a 1-methyl-2-pyrrolidone (NMP) solution and then centrifuged it to obtain heterodimensional nanostructures on the upper layer of the solvent, as shown in Figure 2. In bath sonication, the hydrodynamic forces generated by the increase in pressure and temperature promoted the cleavage of bulk MoS₂ into small particles. Probe sonication broke most of the nanosheets into ultrathin sheets with 0D MoS₂ interspersed. The presence of NMP maintains the stability of the exfoliated MoS₂ dispersion the highest. This method yielded ~2 nm single-layer 0D MoS₂, which slightly agglomerated over time and could become 10 nm after 7 d. Another commonly used solvent is N,N-dimethylformamide (DMF). Wang et al.^[37] placed bulk MoS₂ powder in DMF and successfully prepared 0D MoS₂ of 1.47 nm ± 0.16 nm by a combination of ultrasonic exfoliation for 4 h and gradient centrifugation. Dong et al.^[13] used TBA to insert the interlayer and ultrasonic treatment to assist in the preparation of bulk MoS₂ flakes with a lateral size of ~13 nm. The flakes were uniformly distributed with 0D MoS₂ in a honeycomb lattice atomic structure, a uniform crystal form and a height of 1-2 nm.

Figure 2. Schematic diagram of 0D MoS₂ synthesized by liquid exfoliation method [insets (A,B): TEM images of 0D MoS₂]^[36].

Electrochemical method

The electrochemical method is an environmental-friendly, efficient and condition-controllable method for preparing 0D MoS₂^[38,39]. A direct current (DC) voltage is applied to the electrolyte and free radicals are generated in the solution. The radicals etch MoS₂, thereby triggering its cleavage, leading to further exfoliation of the bulk material that results in monolayer or multilayer 0D MoS₂^[40]. Shrivastava et al.^[41] placed MoS₂ microspheres in a two-electrode system of an electrochemical cell with a 1 mol% lithium bis-trifluoromethylsulphonylimide ionic liquid as the electrolyte. A constant DC voltage of 5 V was applied to the electrode pair and the reaction continued for 3 h. 0D MoS₂ was separated from the supernatant after centrifugation, washed with acetone and dried at room temperature to obtain 0D MoS₂ in powder form. Li et al.^[42] used the electro-Fenton reaction to generate a large number of hydroxyl radicals, which acted on the preparation of 0D MoS₂, as shown in Figure 3. The nanosheets could be completely converted into uniformly distributed single 0D MoS₂ by continuously applying a potential of -0.5 V and bubbling O₂ to the mixed solution containing MoS₂ nanosheets and FeSO₄ for 60 min. The electro-Fenton reaction can be used to easily obtain nanosheets with different nanopores, as well as 0D MoS₂, by controlling the time.

Figure 3. Schematic diagram of 0D MoS₂ synthesized by electrochemical method^[42].

Intercalation method

Intercalation is the most suitable route for the large-scale production of monolayer MoS₂ and 0D MoS₂. By controlling the number of repeated exfoliation steps on the sample by lithium intercalation, monolayer MoS₂ nanosheets with different lateral dimensions can be generated until a large number of 0D MoS₂ monolayers with sufficiently small size can be obtained^[43]. Three important mechanisms contribute to the exfoliation of MoS₂ and the preparation of 0D MoS₂: (1) the intercalation agent intercalates between the layers of the material, which weakens the interlayer force; (2) the intercalation agent reacts violently with water to generate H₂, which further increases the interlayer spacing; (3) coulombic repulsion exists between the layers^[44]. Lithium-containing compounds and some alkali metal compounds can be used as intercalation agents, among which n-butyllithium (n-BuLi) is widely used for the exfoliation of layered materials. Qiao et al.^[43] soaked MoS₂ powder in an n-BuLi solution for a long period, followed by stripping, filtration, washing, sonication and drying. The above process was repeated three times and single-layer 0D MoS₂ was finally prepared. n-BuLi easily leads to the transformation of MoS₂ from the semiconducting 2H phase to the metallic 1T phase during the intercalation process. An et al.^[45] proposed the addition of laser radiation-assisted stripping on the basis of Li⁺ intercalation to induce the 1T phase transition to 5-10 nm 2H phase 0D MoS₂, as shown in Figure 4. It is noteworthy that part of the MoS₂ oxidized spontaneously due to laser ablation.

Figure 4. Schematic diagram of 0D MoS₂ synthesized by intercalation method (A). Digital images of solution change during exfoliation (B). Schematic diagram of experimental setup (C)^[45].

Hydrothermal method

Under the premise of using water as the solvent, by placing the precursor solution in a closed reactor and controlling the reaction time, temperature, pressure and other conditions, 0D MoS₂ with different structures and morphologies can be prepared. Wang et al.^[10] used sodium molybdate and L-cysteine as precursors and reacted them in an autoclave at 200 °C for 36 h to obtain 0D MoS₂ of ~2.5 nm in solution. Ren et al.^[46] converted the sulfur source into dibenzyl disulfide and reacted it with Na₂MoO₄·2H₂O at 220 °C for 18 h to obtain 0D MoS₂ with a uniform distribution and an average size of 3.6 nm, as shown in Figure 5. Mohanty et al.^[47] used ammonium tetrathiomolybdate [(NH₄)₂MoS₄] as the single precursor and hydrazine hydrate (N₂H₄·4H₂O) as the reducing agent and reacted them at 200 °C for 24 h to obtain 0D MoS₂ with an average size of 2.8 nm. This one-step hydrothermal synthesis avoids the inhibition of the performance of 0D MoS₂ by carbon that may be generated in the presence of organic solvents.

Figure 5. Schematic diagram of 0D MoS₂ synthesized by hydrothermal method. Digital image of 0D MoS₂ suspension (A). SEM image of µm-sized 0D MoS₂ (B)^[46]. SEM: Scanning electron microscope.

Top-down methods

Micromechanical cleavage method

Micromechanical cleavage is a traditional top-down method for preparing 2D materials. According to the Van der Waals forces between the material layers, suitable adhesive tapes can be selected for artificial stripping to prepare single- or few-layer 2D materials, including the preparation of 2D MoS₂. Li et al.^[49] exfoliated suitable MoS₂ from bulk crystals using Scotch tape and then contacted these newly peeled flakes on the tape with the target substrate and rubbed them with tools, such as plastic tweezers, to further cut them. After removing the Scotch tape, the obtained optical microscopy and AFM images of single- or few-layer 2D MoS₂ on the substrate are shown in Figure 7A1-8. The height of monolayer MoS₂ is ~0.8 nm, which is consistent with the theoretical value^[50]. Similarly, Ghatak et al.^[51] obtained different layers of 2D MoS₂ by tape stripping, in which schematics of optical microscopy and field-effect transistors are exhibited in Figure 7B1-2. It is obvious that the separation of the E_2g¹ and A_1g peaks for three-, two- and single-layer 2D MoS₂ are 23, 21 and 16-18 cm^-1, respectively, as characterized by Raman spectroscopy [Figure 7B3-4]. 2D MoS₂ obtained from Scotch tape has nearly perfect surfaces and high crystallinity and is usually used as a material for optoelectronic devices in fundamental research. Simultaneously, the high crystallinity of 2D MoS₂, which leads to its low chemical activity, reduces its application prospects in other aspects, such as photocatalysis and adsorption. In addition, the micromechanical cleavage method faces the limitations of small yield and low production efficiency. These problems limit the application of this method to fundamental research at present, and it is difficult to realize the actual production of 2D MoS₂ using this method.

Liquid exfoliation

The basic strategy of liquid exfoliation, similar to chemical mechanical exfoliation, is to disperse the bulk crystals in the liquid phase to reduce the interlayer force and then peel the 2D materials from the bulk and disperse them stably in the solvent by assisted shear or ultrasound. In this method, reducing the interlayer force of MoS₂ is pivotal to preparing high-quality 2D MoS₂. Therefore, significant research has been conducted on the method for effectively reducing the interlayer force of MoS₂ in recent years. According to the different methods to reduce the interlayer force, the liquid exfoliation technology of MoS₂ can be mainly divided into solvent-assisted stripping, surfactant-assisted stripping and ion intercalation stripping.

Solvent-assisted stripping, as one of the most direct methods for stripping layered materials, involves dispersing MoS₂ in an organic solvent and then performing ultrasonic stripping to finally obtain 2D MoS₂ by centrifugation. It has been shown that the energy required for stripping layered minerals is strongly related to the difference between the surface energy of layered minerals and the surface tension of the organic solvents^[52]. In addition, a suitable solvent can stabilize the dispersion of 2D MoS₂ by inhibiting the agglomeration and restacking between nanosheets. In summary, the type of dispersing solvent plays a decisive role in the stripping efficiency of MoS₂ and the commonly used solvents include NMP, dimethyl sulfoxide and DMF^[53]. Li et al.^[54] successfully dispersed massive MoS₂ in an NMP solvent. The high-speed shear effect of the homogenizer was used to overcome the weak Van der Waals forces between the MoS₂ layers in NMP, which reduced the size of the sheet [Figure 8A]. From AFM images of 2D MoS₂ in the supernatant (MoS₂-p) and the histogram of layer number [Figure 8B1-2], the statistical data suggested that the thickness distribution of 2D MoS₂ was wide, while the average layer number was around six. By adding sodium citrate to the NMP suspension, the concentration of 2D MoS₂ in the supernatant (MoS₂-sc) was greatly increased and the dispersion stability was also significantly improved. More importantly, the thickness of the obtained 2D MoS₂ was thinner and the distribution narrower. The average layer number of 2D MoS₂ was about four, of which ~9% were single layers, as shown in Figure 8B3-4. Although the effective stripping of MoS₂ can be achieved by using NMP as a solvent, it is difficult to be popularized due to the toxicity of NMP. Furthermore, the removal of NMP (and other solvents) from 2D MoS₂ is also challenging.

Figure 7. Mechanically exfoliated 2D MoS₂ on 300 nm SiO₂/Si: optical microscopy (A_1-4). AFM images (A_5-8)^[49]. Optical micrograph of a typical MoS₂ device (B₁). Schematic of a single-layer MoS₂ field-effect transistor (B₂). Raman spectrum (B₃). Thickness scan of single-layer MoS₂ (inset: AFM image) (B₄)^[51].

Figure 8. Schematic diagram of liquid-phase shear exfoliation of 2D MoS₂ by a high-speed dispersive homogenizer (A). Typical AFM images (left) and histogram of layer number of 2D MoS₂ calculated from height measurements of AFM images of above 100 flakes (right): MoS₂-p (B_1-2); MoS₂-sc (B_3-4)^[54].

Figure 9. Digital image of MoS₂, surfactant solutions before and after sonication and the zeta potential distribution of as-prepared MoS₂-CTAB and MoS₂-SDS dispersions (A₁). Tapping-mode AFM images (the height profiles of the sheets along the blue line marked on the images are indicated in white), electron microscope images (inset: selected area diffraction pattern) and TEM and HRTEM images of MoS₂-CTAB (A_2-5)^[55]. AFM images of platelets exfoliated in water with Tween 80 and deposited onto mica substrates: h-BN (B₁); MoS₂ (B₂); WS₂ (B₃). Histogram of thickness distribution of the sample in the corresponding AFM image: h-BN (B₄); MoS₂ (B₅); WS₂ (B₆)^[56].

Surfactant-assisted stripping is the other liquid exfoliation method. Small molecular organics, surfactants and polymers have high adsorption energy on the molybdenite surface, which greatly promotes its separation. The surfactants can be divided into ionic and nonionic depending on their structure. Gupta et al.^[55] obtained stable water dispersions of 2D MoS₂ by ultrasonic treatment in the presence of cationic surfactant cetyltrimethylammonium bromide (CTAB) or anionic surfactant sodium dodecyl sulfate (SDS). Observation of the micromorphology of 2D MoS₂ stripped by CTAB revealed that it was highly dispersed and exhibited a defect-free structure and a hexagonal honeycomb shape [Figure 9A3-5]. As shown by the AFM analysis in Figure 9A2, the thickness of 2D MoS₂ was ~1.2 nm, indicating that the nanosheets in the MoS₂ dispersion were composed of one or two layers, irrespective of the surfactant used for exfoliation. Interestingly, the magnitude and distribution of the potential energy of the two dispersions [Figure 9A1] were similar, but their signs were opposite (CTAB was positive and SDS was negative). These charges must have been generated by the interaction between 2D MoS₂ and the surfactant because there were no ionizable groups on the 2D MoS₂. Guardia et al.^[56] used a series of nonionic surfactants, including polyoxyethylene sorbitol monooleate (Tween 80), polyoxyethylene sorbitol trioleate (Tween 85), PVP, polyoxyethylene (4) dodecyl ether (Brij 30) and so on, to effectively produce h-BN [Figure 9B1,4], 2D MoS₂ and WS₂ [Figure 9B3,6] suspensions with long-term stability (several months). It was clear through AFM analysis that the produced 2D MoS₂ consisted of platelets with lateral dimensions from fifty to several hundred nanometers and thicknesses in several nanometers [Figure 9B2,5]. Unfortunately, the 2D MoS₂ dispersion obtained by this method exhibited anti-aggregation properties and its inherent characteristics were degraded to a certain extent due to the loading and modification of the surfactant.

The ions inserted between the layers can significantly reduce the Van der Waals forces between the layers of a layered material, and then the slices can be easily separated by a simple mechanical external force. Therefore, ion intercalation is considered an excellent liquid exfoliation method to improve the exfoliation efficiency of layered materials. For MoS₂, because its layer spacing is small (~0.65 nm), only Lewis or alkali metal bases with a small ionic radius can be inserted into the interlayers. At present, Li⁺ intercalation stripping is the most versatile and effective method for molybdenite exfoliation, as schematically depicted in Figure 10A^[57]. The method mainly includes three steps. Firstly, MoS₂ is immersed in an inert solvent containing n-butyl lithium to promote the insertion of lithium ions into the interlayer of MoS₂. The intercalated MoS₂ is then immersed in pure water and Li_xMoS₂ is hydrolyzed rapidly to produce a large amount of hydrogen, which causes the interlayer expansion of MoS₂. Finally, 2D MoS₂ is produced by ultrasonic-assisted stripping^[58]. This method is low-cost and scalable. However, n-butyl lithium is highly reductive and reacts violently with water, resulting in serious deformation of the 2D MoS₂ structure. In addition, the intercalation reaction time is particularly long, and the insertion amount of Li⁺ in molybdenite cannot be effectively controlled.

Figure 10. Schematic representation of the formation of 2D MoS₂ (A)^[57]. SEM images (flakes were vacuum filtered on an aluminum oxide membrane) and AFM image of 2D MoS₂ (inset: histogram of the number of layers of 2D MoS₂ measured from AFM images) (B_1-2). PL spectra of bulk MoS₂ and 2D MoS₂ (B₃). TEM images of MoS₂ quantum dots (B₄)^[59].

Later, Yeon et al.^[59] introduced a new strategy for exfoliating nanosheets in an environmentally-friendly and inexpensive aqueous solution using Na⁺ instead of Li⁺. The morphology and thickness of 2D MoS₂ exfoliated with 0.1 M Na⁺ are characterized in Figure 10B1-2. The 2D MoS₂ slices are thin enough to clearly observe film holes through them, with most being monolayer or bilayer MoS₂, according to the AFM statistics. The bulk MoS₂ does not show any identifiable emission owing to its indirect band gap. In contrast, the exfoliated 2D MoS₂ has a PL peak near 650 nm, which is similar to that obtained by stripping bulk MoS₂ with Scotch tape. In addition, a strong PL peak is observed at 565 nm [Figure 10B3], which is considered to originate from MoS₂ quantum dots, similar to 0D MoS₂ observed by TEM in Figure 10B4. Similarly, the electrochemically assisted Li⁺ intercalation method has been used to achieve the effective stripping of MoS₂, as described in detail in the later electrochemical stripping section.

In order to achieve green and efficient liquid exfoliation, external forces have also been developed to prepare 2D MoS₂. Usually, grinding and shearing can be used to separate layers from bulk MoS₂, thereby increasing the exfoliation yield of MoS₂ [Figure 11A]. Yu et al.^[60] used pure water as a solvent, separated bulk MoS₂ through the mechanical force between sandpapers in a mechanical sanding machine and dispersed 2D MoS₂ in water by a cell grinder, thereby realizing this stripping process. It is found by SEM in Figure 11b1-3 that the exfoliated 2D MoS₂ is randomly stacked and the thickness of the sheet is several nanometers. Its transverse size is found to be in the range of 500 nm to 5 µm by TEM [Figure 11B4-6], which is consistent with the SEM results. Lee et al.^[61] dispersed a MoS₂ powder using ultrasound in a water/ethanol system and obtained a 2D MoS₂ supernatant by two-step centrifugation. It was shown that almost the smallest size of MoS₂ sheet could be obtained at high centrifugal speed, with a primary height of 3-4 nm and an average height of ~7 nm through characterizing the 2D MoS₂ deposited on a Si/SiO₂ substrate by AFM. Although the stability of the obtained dispersions may not be as stable as in organic solvents, these methods can realize on-site production and avoid the use of organic solvents.

Figure 11. Typical green exfoliation process of 2D MoS₂ (A). SEM (B_1-3) and TEM Images (B_4-6) of 2D MoS₂^[60].

Electrochemical exfoliation

Since MoS₂ has certain semiconducting properties, 2D MoS₂ can be prepared by electrochemical expansion. You et al.^[62] reported an environmentally-friendly electrochemical stripping method for the rapid and large-scale production of 2D MoS₂ at low cost for the first time. Schematic diagrams of the electrochemically expanded MoS₂ device are shown in Figure 12A1-2, where the bulk natural MoS₂ and platinum foil are connected to the two electrodes, respectively, and immersed in a 0.5 M H₂SO₄ aqueous solution. A potentiostat was used to apply a bias potential between MoS₂ and Pt. In the initial stage, a static bias of +1 V was applied to MoS₂ for 10 min, which was then increased to +10 V and kept constant for 30 min to complete the exfoliation process. Finally, 2D MoS₂ could be obtained by placing the expanded MoS₂ in a DMF solution and then stripping the expanded MoS₂ under ultrasonic assistance. As shown in Figure 12B1-2, the average morphology height of exfoliated MoS₂ on the Si/SiO₂ substrate measured by AFM was ~1.8 nm and the transverse size was ~10 µm. In addition, EDS scans showed only Mo, S and other four elements, indicating pure MoS₂ on the Si/SiO₂ substrate. The sharp, obvious vibration peaks in the Raman spectrum indicate that it has high crystallinity [Figure 12B3-4]. By this method, the obtained MoS₂ maintains its semiconducting properties, but its large-scale production remains difficult.

In order to improve the yield of 2D MoS₂, Zeng et al.^[63] developed a method based on Li⁺-assisted electrochemical exfoliation, which could promote the effective stripping of MoS₂. From the flow chart shown in Figure 12C, it could be seen that lithium foil replaced the platinum electrode to become a new anode. When the current was released, Li⁺ was inserted into the interlayer to reduce the Van der Waals forces between the layers of MoS₂. Subsequently, the MoS₂ after Li⁺ intercalation was gradually stripped into highly dispersed 2D MoS₂ under ultrasonic action. Figure 12D1 shows the TEM image of 2D MoS₂, illustrating its high dispersion in the aqueous solution. The electron diffraction pattern in the flat region of the nanosheets [Figure 12D2] and the corresponding HRTEM image [Figure 12D3] show the hexagonal lattice structure with a lattice spacing of 2.7 Å distributed to the (100) crystal plane. The AFM measurement of 2D MoS₂ showed that its thickness was ~1.0 nm, thereby confirming the existence of 2D MoS₂. In addition, the slices of prepared 2D MoS₂ were detected by AFM, with 92% of them found to be monolayers [Figure 12E1-4]. The most obvious advantage of this method is that it cannot only control Li⁺ insertion by controlling the working voltage, but that it also regulates the crystal phase of MoS₂ by adjusting the discharge process of the device.

Figure 12. Stereogram (A₁) and schematic illustration of 2D MoS₂ electrochemical exfoliation experiment setup (A₂). AFM image (B₁), TEM image (inset: higher resolution TEM image) (B₂)^[62], SEM-EDS (B₃) and Raman spectroscopy of 2D MoS₂ (B₄). Electrochemical lithiation process for 2D MoS₂ (C). TEM image (D₁), SAED pattern (D₂) and HRTEM image (D₃) of 2D MoS₂ and AFM images (E_1-3) and height distribution of 2D MoS₂ (E₄)^[63].

Bottom-up methods

Vapor deposition

Both physical vapor deposition (PVD) and chemical vapor deposition (CVD) have been used to prepare 2D MoS₂. In PVD, the material source (solid or liquid) is vaporized into gaseous atoms or molecules or partially ionized into ions under vacuum conditions. Thin films are then deposited on the substrate surface through a low-pressure gas (or plasma) process, which is one of the main surface treatment technologies. Importantly, PVD is based on magnetron sputtering, meaning that extremely thin 2D films can be grown on very large areas of various substrate materials. Moreover, PVD has significant advantages in precisely defining the atomic scale, thickness control and interface cleanliness when processed in an ultra-high vacuum environment.

Nevertheless, the demands for controllable size, thickness, morphology and scalability have led to the dominant usage of CVD. CVD has been recognized as one of the preferred methods for preparing large-sized 2D MoS₂ with high crystal quality and uniform thickness. The main process is summarized as follows. The precursor containing Mo and S is decomposed into corresponding gaseous molecules under high temperature and specific atmospheric conditions, and then the reaction deposition is carried out on the substrate to generate 2D MoS₂. Among them, the most common is atmospheric pressure CVD (APCVD) with molybdenum trioxide (MoO₃) as the precursor. Of course, in addition to MoO₃, the molybdenum source mainly includes MoCl₅ and elemental Mo^[64], while the sulfur source is mainly sulfur powder. Generally, MoO₃ and sulfur powder are placed in a single-zone tube furnace to effectively CVD grow MoS₂ on the substrate in an inert gas atmosphere [Figure 13A]. The reaction proceeds as a step-by-step vulcanization process described by three intermediate reactions^[65]:

Figure 13. CVD system for deposition of 2D MoS₂ (A). TEM images (B_1-2), periodic atomic arrangement (B₃) and SAED pattern of 2D MoS₂ (B₄)^[66]. CVD: Chemical vapor deposition.

2MoO₃ + S → 2MoO₂ + SO₂ (1)

2MoO₂ + 5S → 2MoOS₂+SO₂ (2)

2MoOS₂ + S → 2MoS₂ + SO₂ (3)

Perkgoz et al.^[66] used APCVD to grow monolayer 2D MoS₂ and the deposition duration and temperature effect were systematically changed to better understand MoS₂ deposition and the influence of these parameters on the quality of the monolayer sheets. Figure 13B1-4 displays the TEM/HRTEM and SAED images of 2D MoS₂ to further understand its growth structure under better conditions. It can be seen that 2D MoS₂ with triangular and “chip” shapes were successfully grown, corresponding to the periodic atomic arrangement shown in Figure 13B3. The symmetry of the diffraction spots, as shown in Figure 13B4, indicates that the monolayer sheets were continuous single crystals with no rotational boundary, illustrating the high crystallinity of 2D MoS₂. Although the use of APCVD can achieve grain sizes of up to hundreds of microns of all-round growth, the process is seriously affected by many restrictions, such as overall thickness control, coverage uniformity and repeatability of operation. The variation in the characteristic morphology is related to the relative content of the precursor loaded into the reactor, leading to the consistency of the coverage and feature size remaining a challenge.

Metal-organic/organometallic chemical vapor deposition (MOCVD/OMCVD) is a new technology in which the gas molecules of organometallic compounds are sent to the reaction chamber with a carrier gas and undergo thermal decomposition reactions to form the compound. The first two letters, “MO” or “OM”, refer to metal-organic or organometallic as the precursor for the growth of sheet. Chowdhury et al.^[67] used Mo(CO)₆, a metal-organic compound, and (C₂H₅)₂S as precursors, which were kept in bubblers in APs at 45 °C and room temperature, respectively [Figure 14A]. A large area of 2D MoS₂ was then deposited in a furnace at 850 °C for 1 min through a needle valve, which was uniformly covered and non-polluting on the whole substrate. The Raman spectrum in Figure 14B1 exhibits that there are two different peaks at ~386 and ~406 cm^-1, which are consistent with the vibrations of the Mo-S phonon mode E_2g¹ (in-plane) and A_1g (out-plane), respectively, thereby proving the existence of MoS₂. After vulcanization [Figure 14B2], most of the defect-induced peaks in the growth samples disappeared, indicating that the defect density decreased and the crystal quality improved.

Figure 14. Schematic of MOCVD growth system (A)^[67]. Raman spectrum of MoS₂ films before (B₁) and after (B₂) sulfurization at 850 °C for 30 min (inset: SEM image of a triangular MoS₂ domain following sulfurization). Schematic of 2D MoS₂ synthesis (C). FTIR spectra of Na-DMDTC and Mo-DMDTC (D₁). UV-vis absorbance (at 210 nm) of the mixture of Na-DMDTC and MoCl₅ with varied molar ratios (d₂). Optical image of 2D MoS₂ (D₃). AFM image of 2D MoS₂ marked with a square box in (D₄)^[68].

Liu et al.^[68] designed a bifunctional precursor, molybdenum dimethyldithiocarbamate (Mo-DMDTC), which is a metal-organic compound. The CVD diagram is displayed in Figure 14C. It is clear that Mo-DMDTC decomposes at elevated temperatures, providing a Mo source for the CVD growth of MoS₂. Importantly, the presence of organic compounds reduces the free energy of MoS₂ nucleation on the substrate, namely, organic promoters, which also greatly promote the growth of MoS₂ [Figure 14D1]. The UV-vis spectra [Figure 14D2] show that when the molar ratio of DMDTC and Mo⁵⁺ reached five, the absorbance was close to saturation, indicating that the coordination number of Mo was close to five. A large area of triangular 2D MoS₂ with a length of 20-30 µm, was obtained on a sapphire substrate based on CVD, as shown in Figure 14D3. The thickness scan analysis of AFM suggested that the thickness of triangular MoS₂ was ~0.65 nm, corresponding to monolayer MoS₂ [Figure 14D4].

In order to prepare large-sized and uniform 2D MoS₂ with excellent electrical properties, researchers have subsequently proposed the thermal decomposition of precursors containing both molybdenum and sulfur atoms. Compared with MoO₃ and MoCl₅, the method based on the pyrolysis of (NH₄)₂MoS₄ to grow 2D MoS₂ has the advantages of a single precursor source and large growth window and therefore the potential for large-scale production. Fei et al.^[69] in-situ observed the thermal decomposition of (NH₄)₂MoS₄ and subsequent 2D MoS₂ crystallization behavior by TEM [Figure 15A]. The pattern evolution with increasing temperature, as exhibited in Figure 15B1-4, was clearly observed by SAED. From room temperature (25 °C) to 400 °C, 780 °C and then 900 °C, the pattern displays no diffraction ring, a fuzzy diffraction ring, a sharper bright diffraction ring and then a discrete diffraction ring, respectively. In summary, the crystallization of MoS₂ occurs at 400 °C-900 °C in the vacuum, the crystal size of MoS₂ increases rapidly after 800 °C and MoS₂ finally decomposes at high temperature. The reaction equation of (NH₄)₂MoS₄ pyrolysis to MoS₂ in a N₂ environment is shown as follows:

Figure 15. Schematic side view of experimental setup of in-situ TEM heating stage (A). Evolution of SAED patterns across a wide temperature range (B_1-4): 25 °C (B₁); 400 °C (B₂); 780 °C (B₃); 900 °C (B₄). TEM images showing the size of 2D MoS₂ increasing with increasing temperatures (C_1-4): 800 °C (C₁); 820 °C (C₂); 840 °C (C₃); 850 °C (C₄). Statistical distribution of 2D MoS₂ size as a function of heating temperature from c_1-4 (D₁). Raman spectra of silicone chip, (NH₄)₂MoS₄ precursor and as-grown MoS₂ (D₂)^[69].

Figure 16. SEM images (A_1-3)^[70] of 2D MoS₂ prepared at pH 4.5 (A₁), pH 7.0 (A₂) and pH 9.5 (A₃). SEM images (B_1-6) of MoS₂-160 °C (B_1,4), MoS₂-180 °C (B_2,5) and MoS₂-240 °C (B_3,6) before inoculation^[71]. SEM (C_1-3) and EDS (C₄) spectrum of as-prepared MoS₂^[72]: flowerlike (C₁); nanosheet (C₂); hollow MoS₂ (C₃). SEM image of MoS₂ RNPs (D₁). TEM image of an MoS₂ RNP (the inset images reveal that the edge thickness was ~10 nm) (D₂)^[73]. SEM images (E_1-3) of PW-MoS₂ (E₁), MWP-MoS₂ (E₂) and PP-MoS₂ (E₃)^[74].

(NH₄)₂MoS₄ → 2NH₃ + H₂S + MoS₃ (120 °C-360 °C) (4)

MoS₃ → MoS₂ + S (> 800 °C) (5)

The average grain size of MoS₂ increases with increasing temperature (820-850 °C), i.e., it is 3.2 nm at 820 °C, 6.5 nm at 840°C and 18.5 nm at 850 °C [Figure 15D1]. The same test result can also be observed in the in-situ TEM images of Figure 15C1-4. Figure 15D2 exhibits the in-situ Raman spectra of the fresh chips, including the (NH₄)₂MoS₄ precursor and the grown MoS₂ (annealed at 850 °C for 30 min). Notice that the peak intensity of E¹_2g is over 50% of the A_2g peak, suggesting the horizontal growth of MoS₂. The distance between the two Raman peaks is ~25.5 cm^-1, corresponding to four to five layers of MoS₂. Compared with other preparation methods, the vapor deposition method can be used to prepare single- or few-layer MoS₂ with a large diameter, and the synthesized 2D MoS₂ has the advantage of high crystallinity. However, the operation process of this method is extremely complex and the energy consumption is very high, so it is still unable to realize the mass production of 2D MoS₂.

Hydrothermal/solvothermal method

Hydrothermal/solvothermal syntheses are typical bottom-up methods that are widely used in the synthesis of 2D MoS₂, owing to their simplicity, high efficiency and strong adaptability. Generally, the Mo source (e.g., MoO₃, ammonium molybdate or sodium molybdate) and S source (e.g., sulfur powder, KSCN or thiourea) are dissolved in an aqueous solution/organic solvent at a certain ratio. The solution is then transferred to a sealed reactor and kept at ~200 °C for a certain period (several to more than ten hours), with 2D MoS₂ then obtained after the reaction. According to the different precursor or preparation conditions (temperature, solution pH, and so on), a variety of nanostructured MoS₂ can be easily obtained, including granular, flakes, flower-like microspheres and so on. 2D MoS₂ can be obtained by the hydrothermal treatment of ammonium tetrathiomolybdate at 350 °C in different solution pH values and the SEM images are shown in Figure 16A1-3^[70]. The samples prepared at pH 4.5 are composed of aggregates that are irregularly shaped and interconnected [Figure 16A1], while the samples prepared at pH 7 are composed of particles with different shapes, rather than aggregates [Figure 16A2]. In contrast, the samples prepared at pH 9.5 are flower-like, consisting of petals with a thickness of 10 nm and lengths of 100-200 nm [Figure 16A3]. These petals aggregate to form microspheres of 300-400 nm, with an increase in pH from 4.5 to 9.5, leading to the aggregation of thin nanosheets into flower-like MoS₂.

Furthermore, 2D MoS₂ can also be grown on carbon felt using phosphomolybdic acid and thiourea at different temperatures, as shown by the macroscopic SEM images in Figure 16B1-3 and the corresponding microscopic SEM images in Figure 16B4-6^[71]. The structure of MoS₂-180 °C is a well-assembled multi-faceted nanoflower with a diameter of 200-300 nm, which is composed of crosslinked nanosheets. As the temperature rises to 240 °C, 2D MoS₂ aggregates into larger particles and forms clusters on the surface of the carbon fibers. The three samples all have a 2D layered structure and the stacking structure becomes more obvious with increasing synthesis temperature. In addition, Wu et al.^[72] used CTAB as a surfactant to assist in the preparation of 2D MoS₂ with various shapes under different growth conditions, including hollow, nanosheets and flower-like [Figure 16C1-3]. Figure 16C4 shows that the nanomaterials are composed of Mo and S with a stoichiometric ratio (Mo:S) of 1:2, with no other elements detected by EDS, indicating pure MoS₂.

Interestingly, 2D MoS₂ with a special shape was obtained by the hydrothermal treatment of sodium molybdate dihydrate and thioacetamide at 200 °C for 36 h, as indicated by the SEM and TEM measurements in Figure 16D1-2, which reveal the morphology of the product^[73]. Finally, 2D MoS₂ with a diameter of ~200 nm and an edge thickness of ~10 nm is observed, which is similar to morphology of radar. Furthermore, the solvothermal method is also a simple and efficient method to synthesize 2D MoS₂. Chen et al.^[74] synthesized ultrathin 2D MoS₂ with ammonium molybdate tetrahydrate and thiourea assisted by polyethylene glycol (PEG). 2D MoS₂ prepared with PEG and pure water, as solvents, was labeled as MWP-MoS₂, pure water was PW-MoS₂ and pure PEG was PP-MoS₂. The prepared PW-MoS₂ was clustered and composed of densely stacked nanosheets. MWP-MoS₂ showed a highly folded surface and layered structure. For PP-MoS₂, the agglomeration of nanoparticles was observed and was obviously composed of randomly oriented layers rather than regular stacked layers [Figure 16E1-3].

Template-based method

2D MoS₂ can be anchored on the surfaces of some templates (including sponge, aerogel and carbon matrix) that possess a 3D architecture via the hydrothermal method, CVD or impregnation method^[75,76]. Benefitting from the facile preparation process, the structural and physical properties of 3D MoS₂-based materials can be easily modified by changing the template. Zhu et al.^[77] prepared a 3D-MoS₂ sponge by anchoring MoS₂ nanosheets and graphene oxide (GO) on the skeleton of a sandwich-structured sponge via a simple two-step impregnation method. A blank melamine sponge was dipped into MoS₂ and GO alcohol solutions in sequence to obtain a 3D sponge@MoS₂@GO after a simple heat treatment at 200 °C in an oven for 6 h [Figure 17A]. This material could adsorb organic molecules and provide multidimensional electron transport pathways, as well as exhibit excellent performance in the degradation of aromatic organics. A 3D porous MoS₂/chitosan aerogel with excellent mechanical strength and stability was constructed by anchoring MoS₂ to the surface of the chitosan aerogel^[78]. Benefitting from the superior photocatalytic performance and porous networks, which promote the mass transfer and shorten the diffusion pathway of the substances, the MoS₂/chitosan aerogel can effectively recover gold from a thiosulfate solution via the direct in-situ reduction of Au(I) to Au⁰ under sunlight.

Figure 17. Synthetic pathway of sponge@MoS₂@GO (A) and SEM images of sponge (B), sponge@MoS₂ (C) and sponge@MoS₂@GO (D)^[77]. Schematic of MoS₂/CA hybrid preparation (E)^[79]. Schematic of MoS₂/CNF aerogel fabrication (F) and SEM image (G) and physical appearance (H) of MoS₂/CNF-2-900 aerogel^[84].

In order to improve the application performance of MoS₂ nanosheets in lithium-ion battery anodes and water electrolysis, 2D MoS₂ can be anchored on the skeleton of a conductive 3D matrix, such as 3D porous carbon aerogels (CAs)^[79], carbonized cellulose aerogels^[80], 3D nickel and CAs^[81], graphene foam^[82] and carbon cloth^[83], which avoids the aggregation and restacking of 2D MoS₂ and ensures the fast transport of both electrons and ions during the electrochemical process. A CA was synthesized by Zhang et al.^[79] through the combination of a sol-gel process and high-temperature carbonization and 2D MoS₂ was subsequently decorated on its skeleton via a solvothermal treatment to obtain a MoS₂/CA aerogel [Figure 17B-E]. The resulting composite was carbonized at 900 °C to obtain a lightweight MoS₂/cellulose nanofiber (CNF) aerogel with a hierarchical structure [Figure 17F-H]^[84]. A novel 2D spatial confinement strategy to in-situ fabricate 2D MoS₂ anchored on 3D porous carbon nanosheet networks was demonstrated by Zhou et al.^[85], in which highly crystalline MoS₂ nanosheets with five or fewer layers very homogeneously and tightly lie on the surface of carbon nanosheet walls (with thickness a ≥ 3 nm). The novel 3D architecture composite exhibited an outstanding long-life cycling capability at high rates, i.e., a specific capacity as large as 709 mAh g^-1 was delivered at 2 A g^-1 and maintained at ∼95.2% even after 520 deep charge/discharge cycles.

Self-assembly

3D MoS₂-based materials have been frequently fabricated with the assistance of GO, owing to its self-assembly feature during the reduction process. When a MoS₂ suspension is mixed with a GO suspension, 2D MoS₂ is prone to deposit on the surface of GO. With a decrease in the oxygen-containing group content on GO during thermal or chemical reduction, the hydrophobicity and internal π-π interactions between GO nanosheets are enhanced, resulting in the aggregation and formation of a MoS₂-graphene hydrogel^[86]. Finally, MoS₂-graphene aerogels can be obtained after freeze-drying. 3D MoS₂/graphene aerogels are mainly applied in supercapacitors and lithium/sodium-ion batteries because the constructed 3D architecture facilitates electrolyte ion transport and improves the electrochemical performance through electric double-layer and faradaic pseudo-capacitance^[87-89].

3D hybrid aerogels of MoS₂ nanosheet/nitrogen-doped graphene (3D MoS₂/N-GAs) with a large surface area and high electrical conductivity were fabricated by Liu et al.^[90] using urea as a reductant [Figure 18A]. A 3D free-standing cylinder was obtained after the reaction, in which the MoS₂ nanosheets were uniformly anchored onto graphene sheets with interconnected pores ranging from several nanometers to several micrometers [Figure 18B-D]. When used as a supercapacitor electrode material, this hybrid aerogel exhibited an excellent specific capacitance of 532 F·g^-1 at a current density of 1 A·g^-1 and superior cycling stability of 93.6% capacitance retention after 10000 cycles at 10 A·g^-1, which was significantly higher than those of 2D MoS₂ and nitrogen-doped graphene aerogels alone.

Figure 18. Schematic of the synthesis process of 3D MoS₂/N-GA porous aerogel hybrids (A). SEM images of N-doped graphene aerogel (B) and 3D MoS₂/N-GAs (C-D)^[90]. Schematic of MGAP fabrication process (E) and photographic and SEM images of MGAP (F-G)^[91]. Schematic of the fabrication process of MoS₂/GS hybrid aerogels (H)^[92]. MGAP: MoS₂-anchored graphene aerogel paper.

To reduce the reliance on polymeric binders, conductive additives and metallic current collectors during the electrode preparation process, as well as to assess the true performance of lithium-ion battery anodes, a MoS₂-anchored graphene aerogel paper (MGAP) was synthesized by Lee et al.^[91] via a self-assembly and compression process [Figure 18E]. The fabricated MGAP, with a density of 0.47 g cm^-3 and a surface area of 5.5 m² g^-1, was able to undergo 180° bending [Figure 18F] and MoS₂ was successfully incorporated into the aerogel paper [Figure 18G]. Another method of one-pot freeze-drying self-assembly combined with in-situ thermal decomposition-reduction for 3D MoS₂/graphene aerogel preparation was proposed by Wang et al.^[92], with significant potential for scale-up. A precursor aerogel was first constructed via the self-assembly of (NH₄)₆Mo₇O₂₄ and CH₄N₂S-loaded GO nanosheets. Upon the subsequent in-situ thermal decomposition-reduction treatment process, (NH₄)₆Mo₇O₂₄ and CH₄N₂S were decomposed to MoO₃/NH₃ and NH₃/H₂S, respectively [Figure 18H]. MoO₃ was subsequently in-situ reduced to MoS₂ by reacting with H₂S and GO was reduced to graphene nanosheets, giving rise to the final MoS₂/GS aerogels.

Sol-gel method

3D MoS₂-based materials can also be fabricated via a sol-gel process and subsequent drying to remove the solvent. During the typical process of gelation, the sol-gel transition of MoS₂ nanosheets creates a 3D network by polymeric binders through physical or chemical crosslinking^[93,94]. A method of vacancy-driven gelation to obtain chemically crosslinked hydrogels from defect-rich MoS₂ nanoassemblies and polymeric binder was reported by Jaiswal et al.^[95]. The defects of MoS₂ nanoassemblies acted as active centers for vacancy-driven gelation with a thiol-activated terminal, such as four-arm poly(ethylene glycol)-thiol(PEG-SH) via chemisorption, forming an elastomeric and robust gel [Figure 19A]. The MoS₂ nanoassemblies were completely encapsulated by PEG-SH, endowing a remarkable mechanical resilience for the PEG-SH/2%MoS₂ hydrogel that could be stretched, bended and twisted easily [Figure 19B]. Furthermore, the hydrogel had an interconnected porous network with a pore size of ~10 µm [Figure 19C and D].

Figure 19. Schematic of PEG-SH/MoS₂ hydrogel synthesis mechanism (A) and photographs (B) and SEM images of PEG-SH/MoS₂ hydrogel (C, D)^[95]. Schematic of preparation strategy of 3D MoS₂ aerogel (E)^[96]. Schematic of the growth process of CNF/MoS₂ aerogel (F)^[97]. SEM: Scanning electron microscope; PEG-SH: poly(ethylene glycol)-thiol.

The lack of connecting/bonding sites in the part of functional groups greatly limits the construction of high-quality MoS₂ aerogels. The functionalization of 2D MoS₂ was achieved by Wang et al.^[96] via the self-polymerization of dopamine monomers, rendering MoS₂ nanosheets rich in functional groups, such as catechol, amine and imine. A 3D MoS₂ aerogel could then be facilely constructed under the crosslink of the functionalized MoS₂ nanosheets and chitosan [Figure 19E]. The resulting MoS₂ aerogel could stand on a flexible cedar leaf without changing its original shape, revealing the high porosity and ultralight properties of 3D MoS₂. Furthermore, Mo⁴⁺ was presented in metallic MoS₂, which links with carboxyl (-COOH) and hydroxyl (-OH) groups in the cellulose chains to form crosslinked metal-carboxylate complexes. These complexes allowed the 2D MoS₂ to effectively encapsulate theCNF at the nanoscale, based on which Yang et al.^[97] synthesized an ultralight, highly porous CNF/MoS₂ aerogel with good mechanical strength via a combined sol-gel and freeze-drying process [Figure 19F].

Due to the encapsulation of 2D MoS₂ by polymeric binders, the catalytic activity and conductivity of 3D MoS₂ prepared by the sol-gel method may be weakened, thereby restricting its application in the fields of supercapacitors, lithium-/sodium-ion batteries and electrochemical hydrogen production. Therefore, these 3D MoS₂ aerogels are primarily applied in the field of solar desalination, fire retardants and microwave adsorption^[98] with the utilization of their physical properties.

Point-like defects

Point-like defects in MoS₂ include anionic, cationic and hybrid vacancies, which are generated from the movement of lattice atoms from their original sites, i.e., internal defects. The existence of point-like defects destroys the crystal lattice of MoS₂, leading to the regulation of its chemical properties and electrical configuration^[102]. So far, numerous studies have attempted to produce point-like defects in MoS₂.

Point-like defects in MoS₂ can be induced by the irradiation of particles, i.e., irradiation-induced defects. Normally, particle irradiation is produced from special equipment, such as a tandem accelerator. As illustrated in Figure 20A, when particles with high energy are shot at the MoS₂ surface, the lattice atoms passively release from their original sites to leave point-like defects in the crystal structure^[103]. Importantly, the density of the point-like defects is accurately regulated by controlling the irradiation strength. As shown from the high angle annular dark-field (HAADF) images in Figure 20B, with increasing Ca⁺ irradiation strength, it can be seen that the quantity of point-like defects (the black holes in the images) gradually increases^[104]. Various sources of radiation, including heavy ions, electrons, protons, X-rays, gamma rays, ultraviolet light and infrared irradiation, have been carried out with this technique^[105].

Figure 20. Schematic of inducing defects in MoS₂ by irradiation (A)^[103]. HAADF images of MoS₂ after Ga⁺ irradiation under various strengths (B)^[104].

Heavy ions with high energies are generally accelerated by particle accelerators. Mishra et al.^[106] used a Ga flux beam under ultrahigh vacuum conditions to irradiate MoS₂. As a result, the Raman signals and XPS spectra of the Mo-3d orbital exhibited an obvious shift, revealing that the original atoms in MoS₂ were replaced by Ga atoms. Furthermore, other heavy metals, including uranium^[107], xenon^[108], bismuth^[109] and manganese^[110], have also been adapted to irradiate MoS₂ and successfully induce point-like defects on its surface. In addition, similar procedures to induce point-like defects have been carried out by using other irradiation particles. In conclusion, although the generation of point-like defects in MoS₂ by particle irradiation is a mature technique, it requires specialistic equipment, which limits its popularization.

In addition, the hydrothermal method is another technique to prepare MoS₂ with point-like defects. During the synthetic procedure, the precursors of Mo and S are firstly dissolved in water and transferred into a closed container. Subsequently, MoS₂ crystals form and gradually grow under an environment of high temperature and pressure. It is noteworthy that the microstructure of synthesized MoS₂ can be regulated by adding different molar ratios of the Mo and S precursors. Normally, when the molar ratio of Mo and S atoms breaks the balance of 1:2, point-like defects will form in the MoS₂ structure. MoS₂ samples with different concentrations of point-like defects were synthesized via a hydrothermal method by Chen et al.^[111] and the majorization of the piezoelectric response for gold recovery was realized. Furthermore, according to the report of Jayabal et al.^[112], 1T-MoS₂ samples with point-like defects were prepared using a hydrothermal strategy and exhibited excellent performance in the hydrogen evolution reaction. As a consequence, point-like defects produced by the hydrothermal technique were executed by only using a Teflon-lined stainless steel autoclave, making it a facile method for wide research applications.

Plane-like defects

Compared to the formation of point-like defects, the generation of plane-like defects in MoS₂ requires more energy from the environment to destroy the crystal structure, thus to form a large area of vacancies. As a common method, thermal annealing is an effective method of inducing large-scale edges through the etching of the MoS₂ crystal. By simply undergoing heating, plane-like defects could be induced in MoS₂. Importantly, the degree of defects can be regulated through the selection of reaction atmosphere and temperature. Jia et al.^[113] studied the formation rule and mechanism of edge defects in MoS₂ through thermal treatment. As shown from the AFM images in Figure 21, MoS₂ displays different areas of plane-like defects after thermal treatments under various conditions. A larger etched area is induced at higher temperature and oxygen content in order to achieve good control of plane-like defects in MoS₂. Meanwhile, the shape of the defects oriented to equilateral triangular pits, which was ascribed to the trigonal prismatic 2H phase of MoS₂^[114]. Relevant research has been reported for further understanding the technique mechanism and a similar conclusion was reached^[115,116]. Using this thermal treatment method, numerous properties of MoS₂ have been improved and widely applied in desalination^[117] and adsorption^[99].

Figure 21. AFM images of MoS₂ after thermal treatments under different conditions [Origin MoS₂ (A); 400℃ for 2 h in air (B); 500℃ for 2h in air (C); 500℃ for 3h in N₂ (B)]^[113].

Charge transfer doping

Charge transfer doping is realized by either physical or chemical interactions between dopants and the supporting material, and the dopants originate from gas (O₂ and NO₂), metal or compound molecules. Dopants can attach to MoS₂ via adsorption, deposition and spin and dip coating. The doping degree can be determined by Raman spectroscopy, the PL peak position and intensity charge, the XPS peak position shift, the change in source-drain current-gate voltage response in electrical measurements, Hall measurements and so on^[129]. The activity of the charge transfer direction is governed by the difference at the Fermi level (E_F). The main experimental methods to characterize the doping effect include determining the type and which created by the dopants and supporting materials [Figure 22A]^[121]. If the E_F of the dopants is lower than that of the supporting materials, electrons will be transferred from the supporting materials to the dopants, i.e., p-type doping. In contrast, n-type doping is based on the electron transfer from dopants to supporting materials. Generally speaking, the interaction regarding charge transfer happens on the interface of any adjacent media, like particle ions, surface atoms, molecules and supporting materials^[23,130-132]. The advantage of charge transfer doping is enhancing the electronic behavior of semiconductors whilst avoiding lattice distortion and high mobility transport.

Figure 22. Behavior of charge transfer doping (A)^[121]. Modulation of PL property of MoS₂ via exposure to various gas atmospheres and the change in exciton and non-radiation recombination (B and C)^[137]. Electrical characteristics of K-doped MoS₂ (D-E)^[134].

The main purposes of charge transfer doping are changing the work function of the material and reducing the Schottky barrier between the material and electrode with an improvement in device performance^[133]. It also regulates the type and concentration of carriers in the channel to obtain high-performance devices with electrons (n-type) or holes (p-type) as the main carriers or to construct space charge regions (p-n junction) to provide basic structural elements^[134]. Finally, charge transfer doping can be used to turn the field-effect transistor by constructing a steep in-plane tunneling junction with high heterogeneity.

Gas molecules are often adsorbed on the surface of MoS₂ and then turn the optical and electronic behavior of the supporting materials^[135]. This characterization is applied in gas sensing, in which semiconductor materials are employed as the sensing channel^[136]. The simplest method of gas-phase doping is to directly expose MoS₂ to a target atmosphere. The usual interaction is physical adsorption via weak van der Waals interactions, where desorption is possible. During thermal annealing in a vacuum, the exposure effect will be strengthened. This process can clean unnecessary contaminants and increase the concentration of S vacancies, which promote the adsorption of gas molecules.

Tongay et al.^[137] found that after exposing MoS₂ to various atmospheres, such as O₂, H₂O, ambient and inert gases, the PL intensity in n-type MoS₂ was obviously enhanced because of the depletion of electrons and the suppression of non-radiative recombination [Figure 22B and C]. The PL intensity is much higher than that under a vacuum and increases by 10×, 35× and 100× after exposure to H₂O, O₂ and air, respectively, and the peak shifts from 1.84 to 1.88 eV. However, the PL intensity and peak position remain unchanged when the air is changed to Ar and N₂. When the inert gas contacts the surface of MoS₂, there is no charge transfer due to the stable electronic structure, so it is very difficult to gain and lose electrons, meaning that it cannot play a role in regulating the PL intensity and position^[137]. Later, Nan et al.^[138] demonstrated the reinforced modulation of the PL emission of MoS₂ through chemically bonded O₂ in a mild plasma treatment. Although physical adsorption is a reversible method for quantitatively adjusting the doping concentration through controlling the gas pressure, it cannot meet the requirements for preparing controllable and stable doping materials due to its poor adsorption ability.

Metal and metal compounds^[134], like ion complexes^[139] and oxides^[140,141], can be applied to decorate the surface of MoS₂. In general, evaporation and solution case methods that are compatible with existing patterning techniques are applied for decorating MoS₂. It has been proved that choosing the dopant sources according to their work function can be used to predict the electrical doping behavior. Sarkar et al.^[142] studied the electrical doping of MoS₂ via various noble metals, such as Au, Ag, Pd, Pt, Sc and Y. All the metals with large work functions (44.6 eV) were doped into MoS₂. Of these metals, Sc exhibited only a slight depletion effect, while Y induced electron doping behavior with the lowest work function. This phenomenon demonstrates that the dopants with higher work functions have a tendency to act as p-type dopants.

Potassium has a strong ability to supply electrons for most surfaces due to its low electron affinity^[143]. Fang et al.^[134] first reported an n-type-doped thin layer of MoS₂ based on surface charge transfer by depositing potassium atoms on the surface of MoS₂ via vacuum thermal evaporation and controlling the adsorption amount [Figure 22D]. The transfer curve before and after doping [Figure 22E] shows that the device has a large switching ratio (lon/off > 105) before doping. When the channel was completely evaporated of potassium atoms, the source leakage current increased by five orders of magnitude, close to milliamps. Moreover, the source-drain current was almost saturated, indicating the transfer of electrons from potassium to the surface of MoS₂. The surface change density of potassium-doped MoS₂ reached 1.0 × 10¹³ cm^-2. Although the surface charge density of MoS₂ is greatly improved after potassium doping, the current switching ratio significantly decreases. Furthermore, its poor stability limits its practical application, owing to the facile reaction between potassium and O₂ and H₂O.

In addition to potassium, other materials, such as fullerenes (C60) and MoO₃^[144], have been doped on the surface of MoS₂. Lin et al.^[145] researched the effects of C60 and MoO₃ on the surface charge of MoS₂ via vapor methods. Since the lowest unoccupied electrons of the energy level in C60 is higher than the CB of MoS₂, the electron transfer from MoS₂ to C60 is prevented. Furthermore, the p-type doping effect is not produced and the mobility does not decrease, indicating that C60 can be used as an effective protective film. When MoO₃ is used as the covering film, the threshold voltage increases from 0 to 70 eV, indicating that MoO₃ significantly reduces the charge on the surface of MoS₂ and is a suitable p-type dopant. Cs₂CO₃, as an electron injection layer in organic light-emitting diodes, has the ability to provide electrons and produce a strong n-type doping effect on different organic semiconductor materials. Rai et al.^[146] spin-coated amorphous titanium oxide in a solution-gel state on the surface of MoS₂ and found that the electron mobility was as high as 83 cm²·V^-1·s^-1 after doping at room temperature and the intrinsic mobility at 77 K is as high as 501 cm²·V^-1·s^-1. Simultaneously, they reported that only when titanium is in excess, amorphous titanium oxide with oxygen holes has the doping effect because the extra Ti atoms can generate additional energy levels at the bottom of the CB.

In addition to the above methods for achieving surface charge transfer based on gas-phase physical adsorption, surface evaporation or spin coating, solution immersion is also a simple and effective method to achieve the adsorption of doped molecules on the surface of MoS₂^[147]. Du et al.^[148] soaked MoS₂ in a methanol solution containing amino-rich polyetherimide. The in-plane and contact resistance decreased from 19.99 and 5.06 kΩ·m^-1 to 7.65 and 4.57 kΩ·m^-1, respectively. After doping MoS₂, it also has good air stability.

Charge transfer represents a simple and effective doping method to improve the concentration of surface carriers or change the type of carrier in MoS₂. When n-type MoS₂ is doped with an electron-donating dopant, such as an amino group, the electron concentration of the surface of MoS₂ reaches the limit of electron concentration. When patterned p-type doping is carried out with highly electronegative dopants, such as O and F, homogeneous p-n junctions can also be constructed. Compared with substitutional doping, surface adsorption does not destroy the structure of MoS₂ and only transfers the charge; thus, the structure of MoS₂ is more complete. The experimental method of charge transfer doping has also been expanded from gas physical adsorption to solution-phase adsorption, which needs to be further developed in the future to meet the requirements of large-area, controllable, air stable and other practical applications.

Substitutional doping

Substitutional doping involves the replacement of the Mo and S atoms in MoS₂ with cationic and anionic elements of comparable radii^[149,150]. After doping, the generation of n- or p-type MoS₂ depends on the number of valence electrons of the dopants. Substitutional doping requires significant energy because of the replacement of heteroatoms for Mo or S atoms and the bonding of heteroatoms with Mo and S atoms around. There are two main methods to realize substitutional doping. One is in-situ doping in the growth process. The other is the creation of vacancies by post-processing, with heteroatoms introduced to fill the vacancies.

Both the bottom-up chemical synthesis and top-down stripping of MoS₂ have S atom holes, so that MoS₂ shows n-type characteristics. If MoS₂ can be doped into a p-type semiconductor, a MoS₂ homojunction will be constructed. In the crystal lattice of MoS₂ [Figure 23A], the Mo atoms are sandwiched between the outer chalcogenide layers, which screens the incorporated metal atoms into the substitutional sites. Non-metal dopants tend to occupy the anionic sites at the outer layers due to the length limitation of the ionic bonds. For metal dopants, the radii of the foreign dopant atoms are vital for a favorable position. For light metals (Li and Na), which have small radii, the interlayer distance (0.615 nm) is sufficient for intercalation. However, for heavy dopants with large radii (e.g., Mn, Co, Zn, and so on), substitutional doping is preferred thermodynamically^[149].

Figure 23. Substitutional doping of MoS₂. Crystal structure of MoS₂ (A). Electron density mapping of MoS₂ with Mo substituted by Nb atom (B). Electronic DOS of Re, Ru and Rh (C) and Nb, Zr and Y substitutionally doped MoS₂ (D)^[151]. Structure (E) and electrical properties (F) of Nb-doped MoS₂^[122].

At the cation sites, many transition metal ions with similar radii to Mo atoms can achieve effective substitutional doping into the lattice of MoS₂, which enriches its functionality [Figure 23C and D]. In theory, Nb is a very suitable electron acceptor for p-type doping in MoS₂ because it has fewer valence electrons than Mo^[151]. From the perspectives of the structures of 2H MoS₂ and 2H NbS₂, the lattice parameters are similar. The covalent radii of r(Mo) (130 pm) and r(Nb) (134 pm) are similar. In terms of oxidation valence states, Mo and Nb are in the 4+ state. Based on the above factors, Nb doping does not cause obvious damage to the original structure of MoS₂. Suh et al.^[122] prepared millimeter-grade MoS₂ single crystals through chemical vapor phase transport using iodine as a transport agent under a growth temperature gradient of 1050 °C -936 °C for 500 h. Finally, monolayer Nb-doped MoS₂ was obtained by mechanical stripping. Figure 23B shows a schematic of the Nb-doped MoS₂ structure by chemical vapor transport. After doping, the hole concentration reached 1.8 × 10¹⁴ cm^-2, which is consistent with theoretical calculations that when an Nb atom absorbs an electron, 0.5% Nb atoms can produce a hole density of ~2.8 × 10¹⁴ cm^-2. Finally, the MoS₂ homojunction was constructed^[122]. As seen in the figure of transfer characteristic, the dependence of Nb-doped MoS₂ on the gate decreases and the current increases, presenting p-type doping characteristics [Figure 23E and F]. Different from the above Nb-MoS₂, which was first synthesized in the bulk phase and then obtained as a monolayer by stripping, Laskar et al.^[152] adopted the method for synthesizing Nb-doped MoS₂ with a thickness of 10 nm by CVD. In terms of doping concentration, the concentration of holes in the bulk phase reached 3.1 × 10²⁰ cm^-3 and their mobility was 8.5 cm²·V^-1·s^-1.

In 2016, Gao et al.^[153] successfully prepared Nb- and Re-doped MoS₂ with a monolayer thickness by direct growth in CVD. The obtained concentration of Re was lower than that of Nb because of the difference in formation energy. In order to make sure whether the dopants were at the substitutional site, aberration-corrected Z-contrast scanning transmission electron microscopy (STEM) was used to observe the position of Nb in the single-layer MoS₂. Furthermore, other indirect methods have been used for detecting the substitutional position. Lin et al.^[154] investigated the stability of dopants by high electron beam irradiation. On the surface of MoS₂, it was found that Au migrates to the edges and vacancy sites when it possesses sufficient energy from electron irradiation, while the Nb doping results in enhanced stability. The successful substitutional doping in CVD growth enables the turning of electrical conductance. The n-type dopant (Re) greatly reduced the contact resistance in transistors.

The doping of Fe, Mn and other metal atoms may make MoS₂ magnetic, which provides conditions for the preparation of spin devices. It is predicted that the Curie temperature of MoS₂ can exceed room temperature when the Mn content reaches 10%-15%^[155]. The chemical bond strength between these dopants and S atoms greatly differs from that of the Mo-S bond, so doping with these atoms may produce lattice defects. Zhang et al.^[125] studied the influence of the substrate on the doping of Mn atoms. When a substrate with a flat surface was used, such as graphene, trace doping could be carried out at concentrations of < 2%. Higher concentrations led to competition between MoS₂ and MnS, resulting in structural instability. When using substrates with suspended bonds on the surface, like SiO₂/Si and sapphire, doped Mn atoms caused defects in MoS₂. They used molybdenum trioxide as a molybdenum source, heated it to 725 °C in the center of a tubular furnace and placed decarbonyl manganese and sulfur powder in an argon gas inlet and heated them to 70 °C and 300 °C, respectively, for in-situ doping during growth. The morphology of MoS₂ with SiO₂/Si and sapphire substrates before and after the doping of Mn atoms showed that MoS₂ had smooth edges before doping that became serrated after doping.

Because MoS₂ is a sandwich structure, the Mo atom is located in the middle, while the S atoms are located in the upper and lower layers, so the substitution of S atoms can be achieved either by in-situ growth by mixing the doped atoms with the S source or by hole creation by post-laser doping. The substitutional anion doping tends to happen in the presence of anion vacancies exposed to the outer surface. Various doping elements that are available in the gas phase, like H, O, Cl, N and P, can be adopted to reduce the resistance.

Anion substitutional doping depends on the diffusion of dopants to the vacancy sites. Fixed point doping can be accomplished through a laser-assisted doping method. Li et al.^[156] realized the Se doping of MoS₂ through a CVD method to obtain 2D MoS_2xSe_2(1-x) with a tuned composition. Ma et al.^[157] demonstrated that using Se sources of diselenodiphenyl at about 350°C achieved 21% selenization. Since N-doped MoS₂ can enhance its catalytic hydrogen evolution reaction ability^[158], Qin et al.^[159] prepared gelatinous precursors by dissolving molybdenum pentachloride and thiourea in an ethanol solution with various molar ratios and then placed the gelatinous precursors in a tube furnace. Chen et al.^[160] etched MoS₂ on the surface of sulfur hexafluoride, trifluoromethane, tetrafluoromethane and O₂, thus obtaining p-type F-doped MoS₂ and constructing a MoS₂ homojunction, which is simpler in structure compared with the traditional heterojunction. The lattice constants between the interfaces were exactly the same. In 2016, Nipane et al.^[161] doped both non-degenerate and degenerate p-type conductance into MoS₂ via inductively-coupled PH₃:H₂ plasma treatment. The hole mobility in P-doped MoS₂ reached 8.4 and 137 cm·V^-1·s^-1, respectively, after correcting for the contact resistance.

Substitutional doping is important for achieving control of the magnetic and optical properties of MoS₂ and enriching the diversity of its composition. Because excessive Mo or S atoms in substituted MoS₂ will destroy the stability of its structure, substitutional doping is not the main method to control the electron concentration. The homogenous distribution of impurity atoms can be obtained by central doping in the growth process of MoS₂, while the laser-assisted method can realize the fixed point doping of dopants. The doping concentration can be controlled by adjusting the laser intensity and exposure time. This method provides a new concept for the construction of homojunctions.

Intercalation doping

Unlike substitutional doping, the van der Waals interlayer of MoS₂ allows for foreign ion intercalation^[162] and atoms^[163] and also molecules to enter into the interlayer space. Intercalation is beneficial for the exfoliation of MoS₂ in solution in order to produce novel batteries and electrochemical cells^[63,164]. Intercalation doping can modulate the electronic structure and crystal lattice of the supporting materials, resulting in physical or chemical phenomena, including charge density and superconductivity.

H and alkali metal ions can be very easily intercalated into the intercalation layers due to their small radii^[165]. Furthermore, they induce electronic structure changes and lattice distortion due to the reductive nature by donating electrons to the lattice. At room temperature, the stable crystal structure of MoS₂ is 2H, i.e., six S atoms are coordinated around Mo atoms to form a trigonal prism coordination environment. In this coordination environment, the 4d orbital of the central Mo ion is divided into three groups and the two 4d electrons of the tetravalent Mo ions are arranged in the lowest orbital, which keeps the system with the highest stability. However, after interlayer doping with alkali metals, the coordination environment of the central Mo ion is changed into the octahedral 1T configuration for maintaining the lowest energy electron configuration. This is because the alkali metal injects more electrons into MoS₂ and the 4d orbital is separated into two groups. The newly injected electrons are arranged with the original two electrons in the lowest energy dxy, dyz and dxz parallel rails. Therefore, the intercalation doping of alkali metal ions brings changes in the crystal structure of MoS₂. It was seen that the injection of electrons changes the electrical properties of MoS₂ according to the distribution diagram of state density before and after doping. The Fermi level has the distribution of state density, which makes MoS₂ change from a semiconducting to a metal state^[166].

Alkali metal intercalation is mainly achieved through the immersion of a butyl solution or electrochemical method. Lukowski et al.^[167] synthesized petal-like multilayer MoS₂ with highly exposed edge sites on a graphene substrate by CVD, then soaked the multilayer petal-like MoS₂ in an n-butyl lithium solution for 6-8 h and finally removed it with deionized water to remove the excess Li. Acerce et al.^[168] prepared 1T MoS₂ with a size of 1-5 µm by first making MoS₂ into a single layer and then stacking it. MoS₂ powder was first soaked in an n-butyl lithium solution, filtered through argon for 48 h and the excess butyl lithium solution was washed away with hexane, before finally soaking the Li-intercalated MoS₂ in deionized water and ultrasound for 1 h to wash away the Li interlayer and obtain the single-layer MoS₂.

The zerovalent intercalation of heavy metal elements, such as Co, Cu and so on, can be intercalated into the intercalation layers^[169-171]. This avoids lattice disruption without a charge interaction between the doping atoms and supporting materials. It was demonstrated that various organic molecules with different functional groups can facilitate the origination of new intercalation systems with engineered enhanced functions, including ferromagnetic moments^[172], charge density waves and superconductivity^[173]. Chen et al.^[174] successfully synthesized the intercalation doping of Pt, Ru, Au and Pd in highly reductive 1T’ Li_xMoS₂ by the in-situ reduction of metal ions. In the process, Li⁺ intercalation enlarges the van der Waal gap in 1T’ Li_xMoS₂, which allows the diffusion of highly charged metal ions (PbCl₆^2-, Ru³⁺, AuCl₄^- and Pb²⁺) into the interlayer space. After in-situ reduction, the noble metal particles are precipitated in the interlayer space. Li⁺ and Cl^- are leached out through ion exchange, the zerovalent intercalated MoS₂ is restored to the 2H phase. The zerovalent is kept in the stable 2H structure. The encapsulation of noble metal clusters in MoS₂ is used for stability in catalytic applications. In addition to the above-mentioned ions and elemental atoms with small sizes, large molecules can also be applied as intercalants. The organic molecules as intercalation dopants facilitate the discovery of new intercalation systems and engineering owing to their enormous functional groups.

Composite construction with single elements

Although its special microstructure and proper bandgap structure endow MoS₂ with excellent physical properties and chemical activity, its properties can be further optimized by constructing composites with single elements. Noble metals are the most common elements used to modify MoS₂ for superconductivity and special surface plasmon resonance. Through some techniques, noble metal nanoparticles can be anchored on MoS₂ to form composites^[175]. According to Sun et al.^[176], MoS₂ decorated with Ag⁰ could be successfully prepared by reducing AgNO₃ on MoS₂. As a result, Ag⁰ dramatically facilitated the delivery of photogenerated electrons from MoS₂, leading to an improvement in the photocatalytic efficiency for Cr(VI) reduction. In another study, Ag/MoS₂ was fabricated as a Fenton catalyst. Under the irradiation of light, “hot electrons” generated by the surface plasmon resonance effect of Ag⁰ can facilitate the separation of electron-hole pairs to co-catalyze Fenton reaction, thus improving the degradation of 2,4-dichlorophenol^[177]. As another kind of noble metal, Au is popular in the modification on MoS₂ to regulate the surface structure. Yin et al.^[178] loaded Au nanoparticles on MoS₂ to improve the photocatalytic features. After Au decoration, the photocurrent response increased from 100 to 370 µA cm⁻² at 0.8 V (vs. Ag/AgCl), ascribed to the local electric field generated by Au. As well as Ag and Au, other common noble metals, such as Pt and Pd, have also been used to decorate MoS₂. Luo et al.^[179] adopted an electrodeposition method to import Pt on MoS₂ and it exhibited a comparable HER activity. As reported by Yuwen et al.^[175], they utilized a microwave-assisted method to disperse Pd on MoS₂, leading to an enhanced catalytic activity for methanol oxidation.

Some heavy metal elements have also been adopted to construct composites with MoS₂ in the lubrication and advanced oxidation process fields. Sun et al.^[180] induced Cu atoms into MoS₂ to form a turbulence-like micro-nanostructured composite (Cu/MoS₂) and utilized it as a solid lubricant. With the introduction of Cu, the composite exhibited higher hardness and a lower friction coefficient in the atmosphere than pure MoS₂. The Fenton reaction is considered a promising technique among various advanced oxidation processes. In the report of Lan et al.^[181], isolated protruding Fe atoms were immobilized on MoS₂ surface through a dipping calcination method to prepare a Fenton catalyst (Fe-MoS₂). During the piezo-activation of peroxymonosulfate, the isolated Fe atomic sites play important roles in promoting the separation of e^- and h⁺ of MoS₂, as well as providing Fe²⁺ for peroxymonosulfate decomposition.

Heterojunctions

The photoactivity of MoS₂ can be enhanced through appropriate heterojunction engineering for inhibiting the recombination of induced photoelectrons and holes. According to the contact interface, heterojunctions can be divided into six groups [Figure 24]: traditional (types I, II and III); p-n; Z-scheme; step-scheme; Schottky and surface heterojunctions^[182-184]. The diverse heterojunction structure shows different methods and mechanisms for the transfer of photogenerated carriers.

Figure 24. Schematic of band structure and charge carrier transfer of traditional heterojunction systems (A-C), a p-n heterojunction system (D)^[191], a direct Z-scheme heterojunction system (E)^[194], a step scheme heterojunction system (F)^[201], a Schottky junction system (G)^[204] and a surface heterojunction system (H)^[205].

In type I heterojunction photocatalysts [Figure 24A], both photogenerated electrons and holes flow from the higher conduction and lower valence bands to another semiconductor. Nevertheless, type I heterojunctions cannot easily separate charge carriers, resulting in the accumulation of electrons and holes in one semiconductor^[185]. Type II heterojunctions effectively avoid the accumulation problems [Figure 24B], as the electrons transfer to the semiconductor with a lower CB and holes transfer with a higher VB, thus decreasing the contact and recombination of electron-hole pairs^[186]. Although the structure of type III heterojunctions [Figure 24C] is close to the type II structure, the bandgap of each does not overlap due to the extreme levels of the bands^[187]. In a traditional heterojunction, only type II heterojunctions possess an ideal structure for improving the separation of electron-hole pairs.

Taking the example of CdS/MoS₂^[188], a type I heterojunction, the photogenerated electrons in the CB of CdS transfer to the CB of MoS₂ due to the more positive conduction of MoS₂ than that of CdS. Therefore, photogenerated electron-hole pairs will be separated in CdS and then enhance the catalytic activity. However, simultaneously, the holes transfer from the VB of CdS to that of MoS₂. Both electrons and holes accumulate in MoS₂, resulting in the ineffective separation in the type I CdS/MoS₂ heterojunction. For a type II heterojunction, like MoS₂/g-C₃N₄^[189], the photogenerated electrons transfer from the CB of g-C₃N₄ to the CB of MoS₂, while the holes move in the opposite direction. As a result, electrons and holes are separated and accumulated in MoS₂ and g-C₃N₄, respectively. The separation of charge carriers can significantly reduce the recombination of charges and achieve high catalytic activity.

Because of the diffusion of carriers between semiconductors, there exists an internal electric field on the interface between p- and n-type semiconductors [Figure 24D]^[190,191]. In the electric field, the electrons and holes will be quickly driven to the CB of the n-type semiconductor and the VB of the p-type semiconductor, respectively. Thus, electrons and holes will be finally separated in the presence of the internal electric field in a p-n heterojunction system. For example, Meng et al.^[38] synthesized a p-n heterojunction (p-MoS₂/n-Bi₂WO₆) through a bath sonication method. Electrons on n-Bi₂WO₆ transfer into the p-MoS₂, while holes on p-MoS₂ move in the opposite direction, thus resulting in a built-in electric field forming at the interface. With the illumination, electrons and holes will be separated quickly via the synergetic effect and fast charge recombination will be efficiently inhibited simultaneously. All the heterojunction catalysts show high separation efficiency of photogenerated electron-hole pairs.

There are three types (traditional, all-solid-state and direct Z-scheme) of Z-scheme heterojunctions. In the structures, the charge transfer method is similar but with no intermediate^[192,193]. In a direct Z-scheme catalyst, recombination occurs between weak electrons with less negative CB and holes from the semiconductor with less positive VB [Figure 24E]. As a result, the high oxidation holes and reduction electrons cannot recombine and be maintained for catalytic reactions^[194,195]. This method gives new insights into the enhancement of catalytic activity by consuming the excess electrons and holes for the remaining high redox potential in the system. Xiong et al.^[196] prepared a Z-scheme heterojunction, (BiO)₂CO₃/MoS₂, for the removal of NO 44. Under irradiation, both (BiO)₂CO₃ and MoS₂ generated electron-hole pairs. The photogenerated electrons in the CB of (BiO)₂CO₃ diffused and combined with the holes in the VB of MoS₂. After that, electrons and holes will be separated and then accumulated at the CB of MoS₂ and the VB of (BiO)₂CO₃, respectively. In addition, ternary heterojunctions are also effective architectures for improving the catalytic activity of binary structures. Zhang et al.^[197] synthesized a ternary nanocomposite, TiO₂/g-C₃N₄/MoS₂, which exhibited two type II heterojunctions at the interface of TiO₂/g-C₃N₄ and TiO₂/MoS₂. Photogenerated electrons will transfer from the CB of g-C₃N₄ (or MoS₂) to the CB of TiO₂. Furthermore, the photogenerated holes in the VB of TiO₂ shift to the VB of g-C₃N₄ (or MoS₂). Finally, TiO₂ will receive the most photogenerated electrons, which reduce the O₂ and •O^2- for degradation. In contrast, g-C₃N₄ and MoS₂ collect the main photogenerated holes. According to previous reports, TiO₂/g-C₃N₄ with the addition of MoS₂ has higher catalytic activity than that without MoS₂ because of the existence of a synergetic effect between TiO₂/g-C₃N₄ and TiO₂/MoS₂ heterojunctions. Beyond that, TiO₂/g-C₃N₄/MoS₂ composite not only extends the light absorption ability but also enhances the light utilization rate. Other ternary catalysts with MoS₂ possess advanced performance.

The step-scheme system is composed of two n-type semiconductors with a staggered band structure, which is similar to type II except for the path of charge migration [Figure 24F]^[198-201]. An electric field is generated from semiconductor A with a greater work function and lower Fermi level to the one (semiconductor B) with smaller work function and higher Fermi level because of the spontaneous diffusion of electrons on the interface. In addition, electron donors will bend upward owing to electron depletion, and another conductor does the opposite with downward bending originating from the electron accumulation^[202]. Therefore, the electrons in semiconductor B will combine with oxidative holes from semiconductor A, which is similar to the Z-scheme heterostructure. Schottky heterojunctions are generated with an interaction between a semiconductor and a metal [Figure 24G]^[203,204]. The electrons are transferred from metal to semiconductor, owing to their different Fermi energies. As for surface heterojunctions, they mainly consist of two different crystal facets of a single semiconductor^[205,206]. The mechanism regarding separation of carriers in the surface heterojunction is similar to that in type II [Figure 24H]^[207]. However, the redox potential in surface heterojunctions is lower than that in type II heterojunctions.

Composite construction with other materials

With increased research, scientists have gradually recognized that pure MoS₂ nanosheets do not meet practical requirements because of their poor conductivity, difficult recovery and low chemical stability^[78,112,208]. As a result, various kinds of functional materials have emerged and been researched in recent years as potential hybrids with MoS₂ to solve these challenges.

Polydopamine (PDA) can be synthesized through a self-polymerization method of dopamine by a simple surface coating in an aqueous solution. As a result of its strong surface adhesion, PDA can adhere well onto almost all inorganic and organic materials. Importantly, the stability can be improved after PDA wrapping, according to the literature^[209]. For the enhancement of MoS₂ chemical stability, Wang et al.^[210] coated PDA on MoS₂ to fabricate a MoS₂@PDA composite as a catalyst to recover AuCl₄⁻. As a result, MoS₂@PDA not only exhibits extraordinary performance in catalysis but also shows strong chemical stability for practical applications.

In order to facilitate the separation of MoS₂ from water for further utilization, it is a common strategy to magnetize MoS₂. Sun et al.^[211] constructed Fe₃O₄ on MoS₂ to fabricate a core-shell structured MoS₂@Fe₃O₄ composite [Figure 25A]. As shown in Figure 25B, this material shows stronger magnetism than pure Fe₃O₄ that can be easily separated from water by a magnet. Furthermore, the MoS₂ component plays the original role, such as the properties of adsorption and exhibited extraordinary effect in recovering gold. Furthermore, Wang et al.^[212] fabricated a ternary composite of Fe₃O₄@PDA-MoS₂ nanospheres [Figure 25C]. This material not only provides a convenient way to separate material from water by a magnet, but also possesses enough chemical stability, thereby promoting its application as an adsorbent for Pb²⁺ removal. In summary, such MoS₂-based nanocomposites combine the merits of both MoS₂ nanosheets and functional materials, promoting the applications progress of MoS₂-based nanocomposites in the relevant fields.

Figure 25. TEM image of MoS₂@ Fe₃O₄ composite (A). Magnetic hysteresis loop of Fe₃O₄ nanospheres and MoS₂@Fe₃O₄ composite (B). Fabrication procedure of Fe₃O₄@PDA-MoS₂ nanosphere (C)^[211,212].