Low-dimensional transition-metal dichalcogenides

Agnieszka Kuc

DOI: 10.1039/9781782620112-00001

1 Introduction

Nanomaterials form a field of materials science, which is devoted to the production and properties of systems with at least one dimension at the nanometre scale. If any of the dimensions is restricted, layered 2D materials are formed; if restrictions appear in two dimensions, one obtains 1D polymer-like systems; and finally, if all three dimensions are scaled down to the range of only few nanometres, 0D clusters or nanoflakes are in subject. These considerations are particularly applicable to the case of carbon, where 3D graphite can be exfoliated down to the 2D graphene monolayers (MLs), which in turn can be rolled up to form 1D nanotubes (NTs) or 0D fullerenes (see Fig. 1). Each of these sp2 carbon allotropes exhibits very different physical properties, especially the electronic structure differs significantly between those allotropes. For example, the parabolic dispersion relation in graphite's band structure – resulting in a zero band gap – changes to linear band behaviour in graphene, where it is described by massless Dirac fermions. On the other hand, NTs can either be metallic or semiconducting, depending on the size and chirality. Fullerenes are always insulators with a large finite band gap, independent of size and shape. Among these carbon nanomaterials, graphene research has been developing extremely fast ever after the successful separation from bulk graphite, what led to the Nobel Prize for Novoselov and Geim in 2010.

Low-dimensional nanomaterials are important in many fields of research and technology. Some examples cover silicon-based semi-conductor devices, optical coatings, micro-electromechanical systems, biomedical research, lasers and electro-optics. Recently, they became extremely interesting as building blocks of next-generation devices for (opto-)electronic applications. As modern electronic devices are strongly miniaturized (nanoscale), several problems start to occur. Traditional electronics with silicon-based field-effect transistors (FETs) often suffers from heat dissipation. At this scale, also quantum effects become very important. To overcome problems of silicon-based technology at nanoscale, one could replace it with nanomaterials that perform better at atomic scale.

In the world of 2D materials, graphene has gained enormous attention, especially for its applications in nanoelectronics. High electron mobility, long-distance spin-transport, or exceptional mechanical properties of graphene are very attractive. Graphene has a potential as a spin-conserver system and it is attractive for spintronic applications. However, weak spin–orbit coupling and zero-band gap disregard graphene as switching material in chargeand spin-based transistors.

These difficulties can be overcome in the semiconducting 2D materials. After the discoveries of CNTs and graphene, other layered and corresponding tubular materials have gained considerable attention. The successful methodologies and knowledge gained in the search for graphite monolayers and CNTs have been extended to other inorganic materials. Though graphene is presently a cutting-edge system, it opens up a variety of new possibilities going beyond the limits of its own properties and applications.

Many materials exist in the layered 3D bulk forms, which can be easily confined to lower dimensions resulting in single layers or tubular structures. Among them, the most known are boron nitride, transitionmetal chalcogenides (TMCs), TX2 (T–Mo, W, Nb, Re, Ti, etc.; X–S, Se, Te), halides (Cl, Br, I), or oxides. Layered 3D TMCs of TX2 type have been extensively studied on experimental and theoretical bases for the last 50 years. There is a huge number of theoretical works on various properties of the TMC layered materials reported to date in the literature. Some of the possible elemental compositions of layered TMCs are schematically shown in Fig. 2.

Weak non-covalent interactions between the adjacent sheets and the anisotropic character of TMC structures result in easy shearing of the layers even under high pressure, leading to very good lubricant properties. Other applications, such as catalysis, optoelectronics and photovoltaics, have been proposed and investigated for this family of compounds. However, it was only in 2011, when TMCs have started their renaissance as potential materials for nanoand opto-electronics after seminal works of Nicolosi and co-workers, and Kis and coworkers. The group of Nicolosi have reported that large-area single layers of TMC can be easily produced using liquid exfoliation technique. Using such a single layer TMC, the group of Kis have produced the first FET based on MoS2-ML (see Fig. 3). Pioneering measurements of this MoS2-ML-based device have shown that at room temperature the mobility is about 200 cm2 V s-1, when exfoliated onto the HfO2 substrate, however, it decreases down to the 0.1–10 cm2 V s-1 range if deposited on SiO2. Various electronic devices have been fabricated based on the MoS2-ML, including thin film transistors, logical circuits, amplifiers and photodetectors.

On the other hand, TMC-NTs have been known for about two decades. In 1992 and 1993, Tenne and co-workers have shown that layered WS2 and MoS2 form, in analogy to carbon, inorganic nanotubes and fullerenelike nanoparticles. TMC-NTs can be produced using, for example, chemical vapour transport technique or by high-temperature annealing of the respective metal trisulphides. TMCs-NTs behave as exceptional lubricants. If the MoS2 NTs or nanoonions are added to base grease, the friction coefficient remains low, even at very high loads. Moreover, MoS2 NTs have been used for catalytic conversion of carbon monoxide and hydrogen into methane and water. These findings are quite unexpected, as the fully bonded sulphur atoms in the TMC surfaces are not expected to be chemically active. Their electronic properties are very intriguing, as depending on the chirality, they resemble monolayers or bulk forms.

The promising use of TMC low-dimensional materials as building blocks in nanoelectronic devices calls for detailed investigations of their physical properties. Therefore, in the following, we will try to summarize what is presently known in the field of TMCs. We will review the recent theoretical developments on the properties of low-dimensional TMCs for applications in nano- and opto-electronics. We will compare physical properties, such as electronic structure or lattice dynamics, and show that quantum confinement to single layers or nanotubes causes significant changes in the properties and opens up the possibility for new applications. Further, we will show that these electronic properties could be tuned by external modulators, such as tensile strain or electric fields. Although TMC materials have been widely investigated for about five decades, their role as single-layer systems is new. We will focus on the electronic properties of TMCs from groups 5–7 with the 2H polytype. 0D through 2D systems will be considered as platelets, nanotubes and layers, respectively. Theoretical findings will be compared to the available experimental data. This chapter is organized as follows: in Section 2, the state-of-the-art synthesis methods of 1D and 2D materials are summarized, Section 3 discusses the structural and mechanical properties of TMCs, and Section 4 reports the electronic properties of TMCs and the possibility to tune them in a desired manner.

2 Synthesis methods

2.1 2D Transition-metal dichalcogenide nano-layers and -platelets

2D TMC materials can be presently synthesized using two types of methods: the top-down technique, where the bulk forms are exfoliated into monolayers (MLs), and the bottom-up approach using substrate materials.

The top-down technique includes micromechanical cleavage, also known as the Scotch-tape technique, liquid and chemical exfoliation, intercalation by ionic species, ultrasonication, and others. After successful application to graphite, the micromechanical cleavage has been extended to other inorganic materials, such as MoS2, BN, or perovskites. Thin TMC flakes can be peeled from their bulk crystal structures, attached to the substrates, and identified using similar methods as those developed for graphene (e.g. by optical microscopy). This method produces single-crystal flakes of high-purity and macroscopic continuity, as in the case of graphene. Such flakes can be characterized and utilized for fabrication of individual devices. The size and the thickness of the flakes produced by the Scotchtape technique cannot, however, be easily controlled and monolayers are in great minority among much thicker flakes. Therefore, this technique is not feasible for large-scale production of TMC-MLs for technological applications.

In order to produce large-area TMC-MLs in more controlled way, liquid exfoliation was proven to be very efficient and promising. In 2011, Coleman et al. reported that liquid exfoliation produces few-layer TMC materials. This method is highly scalable, insensitive to air and water, and can be applied generally to other materials, including boron nitride or graphene. It allows production of hybrid dispersion or composites by blending dispersions of different materials.

Ion intercalation, like e.g. ultrasound-promoted hydration of lithiumintercalated compounds, is another effective method, allowing production of single-layer materials. Zeng et al. have recently shown that TMC-MLs can be produced with high yield through a complex lithiation processes. These intercalation methods are known since 1970s and they have been re-discovered in the past few years. The exfoliation by ionic intercalation was advanced in 1980s by Morrison and co-workers, and it typical involves merging of TMC bulk with Li-containing compounds and subsequent exposure to water. Water interacts with lithium to release H2, which in turn separates the layers. This method was successfully used for various TMC materials, including MoS2, SnS2, TiS2, or MoSe2. The main disadvantage of the method is the structural deformation that may affect the electronic or optical properties of the TMC-MLs.

TMCs can also be exfoliated from the parental bulk materials using ultrasonication in selected liquids, such as organic solvents, polymer or surfactant solutions. Several layered crystals were also successfully exfoliated in aqueous solutions of the surfactant sodium cholate using sonication. This procedure results in flakes of few-hundred nanometres in size and can be stabilized against re-aggregation by solvation or steric repulsion of molecules adsorbed from the solvent.

Taking into account the advantages and disadvantages of the above methods, for electronic and photonic applications, the ion-exfoliation is favoured, while the liquid exfoliation technique is preferred for production of composite materials. Application of 2D TMC materials in fields of nanoand opto-electronics requires control over the size and thickness of the nanolayers.

The second type of synthesis methods is the so-called bottom-up technique. Large-scale device fabrication was demonstrated for graphene through chemical vapour deposition (CVD) on metal substrates or epitaxial growth on SiC substrates. Recently, the CVD methods were applied to grow MoS2 thin films on insulating substrates. 3D MoS2 materials can be produced by chemical reactions, CVD, thermal evaporation, etc. These techniques could be easily explored to produce thin-film TMCs. Zhan et al. have produced MLs and few-layer MoS2 by vapour-phase growth of elemental molybdenum on the SiO2 substrate using electron-beam in presence of pure sulphur. Lee et al. have obtained MoS2 MLs using MoO3 and sulphur reactants on reduced graphene oxide or modified SiO2 substrates. The authors have shown that the MoS2 growth was promoted by surface treatment. Furthermore, using sapphire or SiO2/Si substrates, Liu et al. have produced large-area MoS2 MLs through thermolysis of (NH4)2MoS4 and subsequent annealing in sulphur vapour. In many CVD methods, the thickness of the MoS2 sheets depends strongly on the concentration and thickness of the precursors. Different CVD methods are still at relatively early stage of TMC-MLs production, but are also promising to be leading synthesis of thin films of materials other than MoS2. As a matter of fact, first attempts to produce WS2 nanosheets down to 2–3 layers have been performed through the chemical reaction of W18O49 nanorods with CS2 in hexadecylamine solution.

Hydrothermal synthesis at high temperature and pressure from aqua solutions lead to chemical production of MoS2 and MoSe2. Layered transition-metal sulphides or selenides have also been produced at high temperature from reactions of MoO3 or WO3 with SC(NH2)2 or SeC(NH2)2. In this way, good-quality flakes with sizes in nanoand micrometres range are produced, however, not restricted to the ML thickness. Other TMC thin films have been produced through CVD methods, however, the procedures remain challenging.

Planar small MoS2 clusters (platelets/flakes) of single or few-layers have been produced by Besenbacher and co-workers in standard ultrahigh vacuum chamber using high-resolution scanning tunnelling microscopy (HRSTM). In this procedure, two types of substrates have been used, namely Au(111) and the highly-oriented pyrolytic graphite (HOPG). When molybdenum was deposited on Au(111) surface, a self-assembled regular array of Mo islands was formed, which were transformed into MoS2 nanoclusters after the subsequent sulphidation. Therefore, the gold surface acts as a template that allows dispersion of Mo into islands. Following this procedure, the majority of Mo islands can be changed into crystalline MoS2. The HOPG substrate does not support dispersion of Mo, unless defects were formed due to the ion bombardment. Depending on the substrate, different shapes of MoS2 platelets were formed with single-layer triangular clusters favoured in the case of Au(111), while HOPG supported rather hexagon-like truncated shapes of MLs or few-layers stackings.

2.2. 1D Transition-metal dichalcogenide nanotubes

Transition-metal dichalcogenides can form structures that resemble nested carbon nanotubes (1D) or fullerenes (0D). TMC inorganic nanotubes (TMC-NTs) are highly regular and almost defect-free, as opposed to the inorganic fullerenes (TMC-IFs). In this section, we will focus solely on the synthesis methods of tubular forms of TMCs.

Folding the TMC layers into tubular or IF-like forms comes back as early as 1979, when Chianelli and co-workers reported the formation of tubular MoS2 structures and their possible applications in catalysis. At the same time, Sanders observed that MoS2 ML could close-up and encapsulate NiO nanoparticles. This was obtained, when the authors were studying the reactivity of the oxide particles in hydro-desulphurization of petroleum.

NTs and IFs of transition-metal dichalcogenides have been first synthesized and characterized by Tenne and co-workers in early 1990s, shortly after the identification of carbon nanotubes (CNTs) by Iijima et al. Over the past two decades, significant progress has been achieved in the synthesis of NTs from layered materials. The synthesis of TMC-NTs is, most commonly, performed at elevated temperatures. High temperatures help to accelerate the kinetics of the process and induce fluctuations of the planar layers, leading to folding into tubular forms. Various high-energy procedures have been proposed and implemented, however, they mostly lead to the formation of IF nanostructures, thus, unlikely to become useful to growth macroscopic amounts of NTs.

TMC-NTs can be produced employing the techniques similar to those proposed for CNTs with the difference that they involve also the reactions of precursors containing the TMC elements. Initially, only minor amounts of WS2 and MoS2 NTs could be prepared following this strategy, however, improved procedures lead to fabrication of larger quantities of various nanotubes.