Chapter 1: Contemporary Chalcogenide Materials Chemistry Free

Chalcogenides, as the name suggests, are chemical compounds containing at least one chalcogenide anion and at least one electropositive element. Generally, oxides are considered separately from chalcogenides—sulfides, selenides, tellurides (and polonides)—because of their quite different chemical properties. In addition, metal chalcogenides, notably containing selenium or tellurium as the chalcogen species, are less known and have not been systematically studied to the same extent as the corresponding oxides.

Here, we focus on compounds consisting of metal cation(s) and three important Group 16 chalcogens (i.e., sulfur, selenium and tellurium), which are defined as metal chalcogenides. Considering binary metal chalcogenides as an example, the relevant chemical formulae can be written as M _a X _b , where M is a metallic element and X = S, Se, Te. Metal chalcogenides constitute a tremendous number of chemical compounds with a huge library of component elements drawn from across the periodic table. They form with a vast variety of crystal structures and exhibit an extraordinary array of physical and chemical properties that range from intriguing optical, electrical, thermal, and electrochemical properties to novel magnetic, superconducting, and topological properties. Therefore, they have been investigated and utilised in various fields, such as photovoltaics, thermoelectrics and many aspects of solid-state electrochemistry among others.

Chalcogens possess a rich metal chemistry both in the solid state and in molecular compounds, which can be ascribed to their ability to catenate and bond with multiple metal centres. ^1–3 Hence, metal chalcogenides involve numerous compounds that range from straightforward stoichiometric metal-based sulfides, selenides, and tellurides to complex compounds or solid solutions containing diverse metal and/or chalcogen elements with a variety of valence states and proportions. ^3–6 Chalcogens tend to form electron-pair compounds where the chalcogens exhibit the divalent state with two lone pairs of electrons. ¹  

The chemistry of binary metal chalcogenides is so rich because many of the metal and metalloid elements that bond with a specific chalcogen can form several compounds of different stoichiometry and occasionally the series of such compounds (for a given metal) can be extensive. ¹ An example is indium selenide, which exists in the form of InSe, In₂Se₃, In₃Se₄, In₄Se₃, In₆Se₇ (note InSe and In₂Se₃ themselves also form as different phases—polytypes—with different structures). ⁷ Although complicated structures are not infrequent, generally, the majority of binary metal chalcogenides are derived from basic crystal structure types and can be easily described accordingly. ^1,8 Three-dimensional crystal structure types, usually rock salt (NaCl), hexagonal NiAs, zinc blende and wurtzite structures, as well as two-dimensional layered lattice types related to the trigonal CdI₂, constitute the structures of the majority of such chalcogenides. ^1,8 Noteworthy is that simple selenides and tellurides can often share the same crystal structure types as the corresponding sulfide analogues, although important variations can exist, especially for tellurides. ^1,9 For example, while PbX (X = S, Se, Te) compounds each form with the rock salt (NaCl-type) structure, ¹⁰ SnS and SnSe crystallise in an orthorhombic structure whereas SnTe exists with the same rock-salt structure as its lead congener. ¹¹  

Alkali and alkaline earth metal monochalcogenides are colourless, salt-like compounds that are often soluble in water (especially alkali metal monochalcogenides). ¹² Alkali metal chalcogenides usually crystallise in the antifluorite structure, while those of the alkaline earth metals crystallise with the cubic NaCl structure. ¹² By contrast, transition metal chalcogenides possess various stoichiometries and numerous crystal structures, and the most important and widely studied ones are simple stoichiometries, such as metal : chalcogen atomic ratios of 1 : 1 and 1 : 2. ¹² Several metals, principally the early transition metals (e.g., Groups 4–7) can also form corresponding trichalcogenides, such as NbSe₃, ZrSe₃ and ZrTe₃. ^1,12,13

Transition-metal chalcogenides can demonstrate highly covalent and non-ionic behaviours typified by the zinc blend or wurtzite crystal structures in 1 : 1 compounds of the later transition metals (e.g., ZnS). ¹ Many transition metals combine with chalcogens to yield MX₂ dichalcogenides with a relatively precise 1 : 2 stoichiometry, which crystallise in either two-dimensional or three-dimensional structures. ¹ About two thirds of MX₂ chalcogenides form layered structures, especially for the dichalcogenides of nearly all of the early transition metals in Groups 4–7 (with the exception of Mn). ¹ The layered dichalcogenides exhibit a variety of electrical transport properties, which range from insulators (e.g., HfS₂) through semiconductors (e.g., MoS₂) and semi-metals (e.g., WTe₂) to metals (e.g., VSe₂). ¹ The M–X units in layered MX₂ typically adopt either a trigonal prismatic coordination similar to that in MoS₂ or an octahedral coordination as in CdI₂. ^1,12 The strong chemical bonding between metal atoms and chalcogens is in striking contrast to the weak chalcogen–chalcogen bonding between the layers. Generally, transition metal atoms in MX₂ contribute four electrons to form the bonding state; thus, the metals and chalcogens are endowed with the formal charges of +4 and −2, respectively. For example, the disulfides containing early transition metals such as Ti, V, Mo and W possess valence states of 4, bonding with two chalcogen dianions. ¹² By contrast to the above examples, the disulfides containing late transition metals such as Fe, Co and Ni (and including Mn) generally adopt the pyrite (FeS₂, distorted NaCl type) or related marcasite (FeS₂, orthorhombic; space group Pnnm) structural motif. ¹² Both structure types contain the disulfide S₂ ²⁻ anion.

Group 13 metal elements (e.g., Al, Ga, In) can combine with chalcogens to form metal chalcogenides with multiple stoichiometries, ^7,14–16 which is especially notable for indium selenides. ^7,17 For example, with increasing temperature, In₂Se₃ transforms from the room-temperature form, α-In₂Se₃  ^18,19 to the β-In₂Se₃, ²⁰ γ-In₂Se₃  ^18,21 and δ-In₂Se₃  ^18,21 phases consecutively. Specifically, α-In₂Se₃ itself can crystallise in both hexagonal (2H) and rhombohedral (3R) layered structures, with different stacking of Se–In–Se–In–Se layers in which In is both tetrahedrally and octahedrally coordinated by Se. ^18,19 In₄Se₃ is also layered but has a very different crystal structure to α-In₂Se₃ in which (In₃)⁵⁺ clusters are bonded to Se ions along the bc plane, while the remainder of the In atoms form quasi one-dimensional chains. ^22,23 The relative newcomer, In₃Se₄, (synthesised via solvothermal methods) also crystallises with a layered rhombohedral structure, here consisting of Se–In–Se–In–Se–In–Se layers and lattice parameters of a = 3.964(2) Å and c = 39.59(2) Å. ^24–26 The chalcogenides of Group 13 metals have found applications in thermoelectric conversion devices, ^7,23,27,28 photodetectors, ^29–31 Li-ion batteries ³² and in phase-change memory. ³³  

Group 14 and 15 elements bond with chalcogens to form MX (M = Ge, Sn, Pb, X = S, Se, Te), MX₂ (M = Ge, Sn, X = S, Se) and M₂X₃ (M = Sb, Bi, X = S, Se, Te) compounds. Specifically, (i) GeS, GeSe, SnS, and SnSe crystallise in orthorhombic layered structures at room temperature, ³⁴ while PbX (X = S, Se, Te) compounds possess a rock salt structure; ⁸ (ii) Ge, Sn-based MX₂ chalcogenides form layered crystal structures, containing X–M–X covalently bonded layers separated in the third dimension by van der Waals forces; ^35–37 (iii) Bi₂Se₃, Bi₂Te₃, and Sb₂Te₃ crystallise in rhombohedral, layered structures, while Sb₂S₃, Sb₂Se₃, and Bi₂S₃ have orthorhombic, non-layered structures. ^38–42 The Group 14 and 15 chalcogenides are typically semiconductors with a diverse range of band gaps, which demonstrate potential applications in thermoelectric conversion and photodetectors as well as solar cells and lithium/sodium-ion batteries. ^1,43,44

Ternary metal chalcogenides, containing two different metallic elements or two types of chalcogens, comprise numerous classes of materials. Taking AB₂X₄ (A, B = metallic elements, X = S, Se, Te) as one exemplar stoichiometry, these chalcogenides adopt a variety of structures, exhibit intriguing properties and find widespread applications. ^45–47 Particularly, chalcogenide spinels (AB₂X₄), such as chalcochromites MCr₂X₄ (M = Ba, Cd, Co, Zn, Fe, Cu, Hg, Mn, Ni), thiocobaltite CuCo₂S₄, thiorhodites M _x Rh_3−x S₄ (M = Cu, Co, Fe), and thioaluminates MAl₂S₄ (M = Zn, Cr) often combine a distinctive tranche of semiconducting, magnetic and optical properties. ⁴⁸ Among other ternary chalcogenides of note, CuMX₂ (M = Ga, In; X = S, Se, Te) chalcopyrites are well-known semiconductors with high absorption coefficients, which have strong potential for utilisation as light absorbing materials in photovoltaic devices. ^49–51 Cu₂SnSe₃, a typical ternary diamond-like compound, ⁵² has been regarded as a good thermoelectric material, ⁵³ while CuAgSe ⁵⁴ and Cu₅FeS₄ (bornite) ^55,56 also behave as promising thermoelectric materials, principally arising from inherently low lattice thermal conductivity. Considering examples of further compositional complexity, quaternary metal chalcogenides, such as Cu₂ABX₄ (A = Zn, Mn, Cd, Co, Fe, Hg; B = Sn, Ge; X = S, Se), have attracted intensive attention due to their promising applications in both photovoltaics and thermoelectrics. ^1,44,57–59

Aside from the many stoichiometric chalcogenides such as those highlighted above, solid solutions are common among structurally-related chalcogenide materials and members of such series (such as Bi₂Se₃–Bi₂Te₃ and Sb₂Te₃–Bi₂Te₃) ^60,61 may achieve extraordinary properties that do not exist in either of the individual binary end members. ¹ Combinations of different metal chalcogenides can be of significant scientific and technological value since their functional and mechanical properties can be modulated accordingly over a wide range through compositional regulation. ¹  

One of the important characteristics of metal chalcogenides is that they can exist in various structural configurations, such as amorphous, crystalline and nanocrystalline forms. ¹² Obviously, atomic radii, metal-to-chalcogen proportion, and valence electron concentration play decisive roles in determining the structural types and features. ¹ Here, we focus on the structural characteristics of typical metal chalcogenides with layered structures.

The development of layered metal chalcogenides has a long-standing history. For example, the crystal structure of the well-known material MoS₂ was first determined by Dickinson and Pauling as early as 1923. ⁶² Much later, Feldman et al. first synthesised MoS₂ as nanotubes (in large quantities with uniform morphology) by utilising the gas-phase reaction between MoO_3−x and H₂S in a reducing atmosphere at elevated temperatures (800–950 °C). ⁶³ Various nanostructured layered metal chalcogenides have subsequently appeared within researchers' sights, and the momentum of this research built considerably following the successful exfoliation of two-dimensional monolayer graphene by Novoselov et al. in 2004. ⁶⁴ These two-dimensional chalcogenides, possessing finite band gaps in a departure from graphene's zero band gap, are currently at the forefront of research in the disciplines of physics, chemistry, and materials science. ⁶⁵ Layered metal chalcogenides are epitomised by stacking stratiform structures composed of covalently bonded two-dimensional planes supported by interplanar van der Waals forces. These materials have rich chemical compositions and exhibit various exotic, anisotropic physicochemical phenomena such as tuneable energy gaps, superconductivity, magnetism, topological properties and charge density waves. ^66–68 In addition, layered chalcogenides can be exfoliated relatively easily into materials of N layers (N ≥ 1) with different physical and chemical characteristics. ^69–71 For example, several layered metal chalcogenides (e.g., MoS₂, MoSe₂, WS₂ and WSe₂) exhibit a transformation from indirect-band gap semiconductors in the bulk to direct-band gap semiconductors in monolayer form. ⁶⁵ Such exfoliated chalcogenide layers can be widely applied in the fields of electronics, photoelectronics, energy and catalysis. ^69,70

The electronic structures and functional properties of layered metal chalcogenides can be manipulated through various strategies. ^13,72 Other unknown and novel properties are also expected to be discovered by following these strategies. ^68,73 Generally, the ‘standard’ approaches that have been developed over time for modifying bulk materials fall into the following categories: (i) defect engineering, the manipulation of zero-dimensional (point), one-dimensional (dislocation) and two-dimensional (stacking) defects; ^74–76 (ii) intercalation, the insertion of guest species such as metal ions and/or molecules between the layers of metal chalcogenide hosts; ^77,78 (iii) substitution, the chemical substitution of constituent elements with other (extrinsic) elements, including cation substitution and anion substitution; ^79,80 (iv) layer-number tuning, the exfoliation of bulk layered chalcogenides into specific numbers of layers, i.e., single, double, and multiple layer(s); ^70,71 and (v) interfacial interaction, the formation of hetero-interfaces and deposition of chalcogenides on a substrate surface. ^81,82

Layered metal chalcogenides present a great variety of compositions and crystal structures, as exemplified by the dichalcogenides of transition and Group 13–15 metals. Generally, the structural configuration of layered chalcogenides is that the bonding of metal cations and chalcogens leads to the formation of independent monolayers, and the layers are stacked to form stratiform structures via van der Waals interactions. For different layered chalcogenides, the structural configuration of individual monolayers and the stacking pattern between layers can be different.

Here, we will take MX₂ (M = transition metal elements, X = S, Se, Te) as examples to illustrate the variance in crystal structure in detail. Each independent monolayer of MX₂ has a ‘sandwich’ structure consisting of upper and lower layers of X atoms and a middle layer of M atoms in the form of X–M–X stacking. ^65,83 The basic MX₂ ‘sandwich’ has two different types: (i) MX₆ trigonal prisms, in which the upper and lower chalcogen layers both form equilateral triangles in the ab plane, which are aligned directly above one-another along the c axis with the transition metal between them (i.e., located in the centres of each trigonal prism); (ii) MX₆ octahedra, where the upper chalcogen triangle is rotated by 180° in the ab plane with respect to the lower (and again M is located between the respective triangles). ^83,84 We use three letters to represent the positions of the three layers in the ‘sandwich’ structure from top to bottom, with capital letters representing chalcogens and lowercase letters representing transition metals. Hence, the trigonal prismatic and the octahedral coordination environments (layers) can be expressed as AbA (upper and lower X layers are similarly aligned) and AbC (upper and lower X layers are differently aligned), respectively.

The adjacent MX₂ layers are connected in the third dimension via weak van der Waals interactions. Due to the different ‘sandwich’ units and the ways in which the layers are stacked, MX₂ exist in a variety of related structural types. These structures are classified as follows: using Arabic numbers to represent the number of independent monolayers contained within a unit cell, and the letters T, H and R to denote either the trigonal, hexagonal, or rhombohedral crystal systems, respectively, of the unit cell. ^65,83 By considering this formalism, we can introduce the three major crystal structure types, that is, 1T, 2H and 3R (Figure 1.1). The 1T type structure contains ‘sandwich’ layers with octahedral configuration that stack with the sequence …AbC AbC…, possessing a trigonal rotational symmetry. ^65,83 As for the 2H type, the unit cell contains two ‘sandwich’ layers with trigonal prismatic configuration; thus in each ‘sandwich’ the X atoms in the lower layer are vertically aligned with the X atoms in the upper layer, leading to a stacking sequence of …AbA BaB AbA BaB… with a hexagonal rotational symmetry. ^65,83 The configuration of each ‘sandwich’ layer in the 3R type is the same as that in 2H type, (i.e., trigonal prismatic) with the main distinction between structures being the stacking pattern of the ‘sandwich’ layers, which is …AbA CaC BcB AbA CaC BcB… in the 3R type structure, which possess a rhombohedral symmetry. ^65,83

Since the discovery of carbon nanotubes (CNTs) in 1991, ⁸⁵ nanomaterials have sparked tremendous interest, which can be ascribed to their peculiar physicochemical properties and great potential in advanced miniaturised functional devices. The term ‘nanostructure’ requires that at least one dimension of a material falls in the nanoscale, thus comprising zero-dimensional nanoparticles (e.g., quantum dots), one-dimensional nanostructures (e.g., nanowires, nanotubes, nanorods, nanobelts, etc.), two-dimensional nanostructures (e.g., nanosheets) and three-dimensional bulk structures consisting of nanograins and/or incorporating nano-inclusions. The main differences between nanomaterials and bulk materials are the high specific surface area and high surface-to-volume ratios of the former as compared with the latter. With a typical dimension spanning from 1–100 nm, nanomaterials bridge the gap between molecules and bulk materials (such as crystalline solids). Aside from a high specific surface area, materials in this size range exhibit unique and fascinating features compared with their bulk analogues, such as a reduced melting point, improved catalytic activity, enhanced electron transport and a modified band structure. ^43,44,86

The two design strategies for fabricating nanostructured materials are often referred to as ‘top-down’ and ‘bottom-up’ methods. Both are of great significance to nanotechnology and modern industry and each is highly applicable to the synthesis of nanostructured chalcogenides. Top-down methods generally obtain desired nanostructures through breaking down bulk materials via mechanical, physical or chemical approaches. The major drawback of these methods is the difficulty in maintaining order in extended structures and retaining structural periodicity. Moreover, surface structure imperfections may be introduced into materials in this way. ^86–88 The defects that are introduced by top-down approaches can exert a significant influence on the physical properties and surface chemistry of nanomaterials, because the surface-to-volume ratio of nanomaterials is crucial. As the name suggests, bottom-up methods involve building materials from the bottom, such as atom-by-atom, molecule-by-molecule, and cluster-by-cluster. This procedure is expected to produce nanomaterials with more homogeneous compositions and controllable morphologies, while permitting particle size to be tuned and the surface to be functionalised. A drawback, however, can be the necessary utilisation of toxic precursors or the employment of reagents with difficult or restricted handling requirements. ^86,89–91

Since the first report of the bottom-up syntheses of MoS₂ and WS₂ nanotubes in the 1990s, ^63,92,93 other nanostructured metal chalcogenides have attracted tremendous attention due to their distinctive morphological characteristics and promising physicochemical properties. Among chalcogenide nanostructures, nanocrystals or quantum dots of semiconductors consisting of chalcogens and Group 12 and 14 elements, such as CdX, ZnX and PbX (X = S, Se, Te), offer unusual and frequently useful functional properties. ^12,94 A few Group 4, 5 and 6 metal dichalcogenides, such as HfS₂, ZrS₂, NbS₂, NbSe₂, MoS₂, MoSe₂, WS₂ and WSe₂, possess layered structures that are beneficial to nanotube formation. ⁹⁵ In addition, two-dimensional layer-structured metal chalcogenides, such as Bi₂Se₃ and Bi₂Te₃, normally grow in a thermodynamically favoured manner, which can grow into a nanoplate/nanosheet form even without the assistance of shape-controlling agents. ^43,96

Nanostructured metal chalcogenides and their heterostructures have aroused extensive interest in multifarious applications, such as thermoelectric devices, solar cells, lithium-ion batteries, supercapacitors, catalysis, fuel cells, sensors, light-emitting diodes, and phase-change memory materials. ^1,43,44,97 The properties of nanostructured metal chalcogenides can be intentionally regulated depending on their specific applications. For example, Bi₂Te₃-based thermoelectric materials consisting of nanoscale grains and nano-inclusions display reduced lattice thermal conductivity, due to strengthened phonon scattering. ^98–100 Many nanostructured metal chalcogenides (e.g., CuS) have become popular options for solar absorber materials. ¹⁰¹ Nanostructured transition metal chalcogenides are hailed as promising non-noble metal catalysts for oxygen reduction reaction and are expected to be effective replacements for precious Pt-based electrocatalysts. ^102,103 The use of germanium chalcogenide nanostructures (e.g., GeSb₂Te₄) for data storage devices could hopefully pave the way for future large-scale applications ^104,105 and finally CdSe can produce any colour across the visible spectrum by adjusting the size of its constituent particles. ¹⁰⁶ Hence it can be seen, just from this selection of examples, that many efforts can be made to optimise the fundamental properties of metal chalcogenides by modifying their size, morphology, crystallinity, and defect structure through nanostructuring.

In the 21st century, with the exploration of several novel physical phenomena, various new metal chalcogenide systems have been discovered and corresponding application fields have been further expanded. ^107–109 A great diversity of multiple-element combinations and stoichiometries have been demonstrated to exist in solid state metal chalcogenides. ¹ In addition, non-stoichiometric phases are abundant for metal chalcogenides, especially for those containing transition metals, where variable valences are favoured and electronegativity differences are minimal. ¹ In this section, we concentrate on two archetypal metal chalcogenide materials systems, MoS₂ and Bi₂Te₃, for the reason that the diverse chemical and physical properties of these layered chalcogenides have enabled their successful application across different fields. ^{38,61,65,110–112} Here, we focus on their intrinsic structure and property characteristics and briefly introduce their corresponding fields of application. It should be noted that there are a wide variety of other chalcogenides with many other attractive properties and applications, but these cannot be covered here due to limited space and scope.

Molybdenum sulfide (MoS₂), with a layered crystal structure and available in nature as the mineral molybdenite, is perhaps the most studied transition metal dichalcogenide. It has been investigated in the fields of dry lubrication, ¹¹³ photovoltaics, ¹¹⁴ catalysis ¹¹⁵ and batteries ¹¹⁶ since the 1960s. Recently, few-layer MoS₂ has attracted considerable attention due to its semiconducting characteristics and the potential for its application in electronics and optoelectronics. ⁶⁵ Different from graphene, which has a single plane of atoms per sheet, mono-layer MoS₂, comprises a plane of Mo atoms sandwiched between two planes of covalently bonded S atoms, lacks the ‘wrinkled’ morphology of graphene and characteristically forms smoother, more planar sheets at the nanoscale. ¹¹⁷ Each single layer of MoS₂ is about 6.5 Å thick. ^65,118 The crystal structure of MoS₂ is composed of stacked S–Mo–S layers held together by weak van der Waals forces ^65,118 and the van der Waals gap has been measured to be about 3.49 Å. ¹¹⁹ Within the S–Mo–S layer, the electronegativity difference between Mo and S results in covalent bonds that are partially polarised. Mo relinquishes (primarily d-band) valence electrons to S, and is left in an oxidation state of +4, while the oxidation state of the S anions becomes −2. ¹²⁰  

Bulk MoS₂ is composed of vertically stacked, weakly interacting S–Mo–S layers, which can be easily sheared by mechanical cleavage and liquid exfoliation. ¹²¹ Hence, single- or few-layer MoS₂ can be produced easily and the sheets studied individually. ¹¹⁰ Previous studies revealed that MoS₂ is a polytypic material based on the coordination of S atoms with Mo atoms within the sheets and the manner in which the sheets are assembled, ¹²² MoS₂ is categorised into three fundamental structural polytypes, that is, 1T-MoS₂, 2H-MoS₂, and 3R-MoS₂, ^65,67 the space groups of which are P3̄m1, P6₃/mmc, and R3m, respectively. ^123,124 Although the three polytypes have different lattice parameters and cell symmetries, the Mo–S bond lengths are approximately 2.4 Å for all of them. ¹²⁵  

Among the three polymorphs of MoS₂, the 2H and 3R polytypes are found to be naturally occurring, while the metastable 1T polytype is a purely synthetic phase. ¹²⁶ Density functional theory calculations ¹²⁷ and X-ray diffraction experiments ¹²⁸ have indicated that in fact the 1T structure of MoS₂ should indeed be thermodynamically unstable. 2H- and 3R-MoS₂ possess Mo in trigonal prismatic coordination, while 1T-MoS₂ contains octahedrally coordinated Mo atoms. ¹²⁶ The alternative stacking arrangements of the trigonal prismatic Mo–S layers lead to the structural difference between 2H- and 3R-MoS₂. Despite some small differences in band structures, 2H- and 3R-MoS₂ have nearly indistinguishable properties, such as n-type semiconducting properties and intrinsic diamagnetism. ¹²⁹ On the contrary, 1T-MoS₂ is paramagnetic and metallic, consistent with the change in coordination geometry of the Mo atoms. ¹³⁰ As expected, the complete filling of orbitals (bands) results in semiconducting behaviour while partial filling leads to metallic behaviour. ¹³¹ Therefore, as the orbital occupation varies, the electronic properties of MoS₂ gradually change from metal through semiconductor to topological insulator, which provides a diversity of chemical and physical properties for wide varieties of applications. ^67,131

For 2H-MoS₂, the stacking of the trigonal prismatic Mo–S layers yields a thermodynamically-stable structure, ¹³² which is the most ubiquitous polytype and its widespread application benefits from its plenitude in the Earth's crust. ¹³³ In the majority of studies, it is the 2H-MoS₂ polytype that has been utilised as a precursor to prepare mono- or few-layer MoS₂ nanosheets. ⁶⁷ Given the stability of the 2H polytype, it is also worth noting that 3R- and 1T-MoS₂ can easily transform into the 2H form. ¹²⁹ For example, 3R-MoS₂ can be converted to the 2H phase through heating, while 1T-MoS₂ can be converted to the 2H phase by elemental doping, annealing or by applying strain. ^134–138 Interestingly, the semiconducting 2H and metallic 1T phases can coexist in chemically exfoliated MoS₂ mono-layers, which was observed utilizing scanning transmission electron microscopy (STEM). ¹³⁹  

Basically, due to the changes in degree of quantum confinement, interlayer coupling and symmetry elements, exfoliated thin-layer MoS₂ nanosheets exhibit dramatically distinct physicochemical properties from bulk MoS₂. ⁸⁴ For example, bulk MoS₂ has an indirect band gap of ∼1.3 eV (Figure 1.2), a valence-band maximum (VBM) that is highly dependent on the interlayer interactions and is black in colour due to the absorption of all visible light. ^140,141 In contrast, mono-layer MoS₂ is a direct band gap semiconductor with a conduction-band minimum (CBM) and VBM coinciding at the K-point (Figure 1.2), ⁸⁴ meanwhile its band gap increases from ∼1.3 eV to ∼1.9 eV due to the quantum confinement effect. Accordingly, a strong photoluminescence, which is absent in bulk MoS₂, emerges in single-layer 2H-MoS₂. ¹⁴² In addition, the colour of MoS₂ samples, prepared from MoS₂ powders in N-methyl-2-pyrrolidone by sonication and centrifugation, varies from black to light-yellow with decreasing layer number and size, indicating the widening of the band gap. ¹⁴³ Moreover, the crossover from an indirect to a direct band gap semiconductor accounts for the luminescence enhancement in single-layer MoS₂. ⁶⁷  

Because of the strong in-plane Mo–S bonding and weak out-of-plane van der Waals interaction, the mechanical and functional properties of MoS₂ are highly anisotropic. ¹⁴⁴ For example, thermal expansion in the direction perpendicular to the basal plane (i.e., parallel to the c-axis) is more substantial compared with that of the corresponding planar direction(s), while the thermal conductivity is orders of magnitude lower in the c-direction, normal to the basal ab plane. ¹²⁰ Moreover, the crystal structure of MoS₂ offers two-dimensional penetrable channels for electron transport, whereas the movement of electrons in the perpendicular direction is hampered by an interlayer potential difference. ^65,118 Therefore, the electrical conductivity within layers is higher than that across adjacent layers. Thus, decreasing the layer numbers contributes to less constrained electron transport and improved electrical conductivity. Additionally, MoS₂ possesses structural defects, such as point defects (vacancies), grain boundaries and edges. ¹⁴⁵ These defects offer the feasibility of surface modification and functionalisation, which increase the options for the further manipulation of the physicochemical properties of MoS₂. ¹⁴⁵ For example, functionalisation of MoS₂ can be achieved via the step-by-step covalent assembly of lipoic acid and fluorescein isothiocyanate, contributing to the simultaneous tailoring of fluorescent and electronic properties. ¹⁴⁶  

MoS₂ is widely explored in various applications due to its unusual combination of physical properties such as high carrier mobility and tuneable band structure. ^147,148 Altering the morphology provides another route by which the physical properties of MoS₂ can be modified; particularly relevant are variations in structural stability, band gap and optical properties. ¹²⁹ In terms of mechanical properties, MoS₂ has an ultrahigh mechanical strength and a larger Young's modulus than steel. ^149,150 Single-layer MoS₂ is marginally stronger than its bulk counterpart. ¹⁵⁰ In addition, MoS₂ layers exhibit remarkably high mechanical flexibility. For example, single-layer MoS₂ can be deformed by an effective strain of up to 11% without fracture, ¹⁵⁰ such that fabricated MoS₂ thin-film transistors can be bent to a curvature radius of 0.75 mm without obvious damage or denigration of electrical properties. ¹⁵¹ These features emphasise the role of MoS₂ as a promising material for flexible electronics.

In the tribology field, MoS₂ has long been recognised as an exceptional lubricant due to the relative inertness of sulfur basal planes and the weak van der Waals forces between S–Mo–S layers. ¹⁵² It is a favoured dry lubricant for aerospace applications; a status which can be ascribed to its outstanding lubricity under high load and in vacuo. ¹⁵³ In terms of energy storage, MoS₂ is a promising electrode material for lithium-ion batteries, where the rather mediocre intermediate insertion voltage (∼2 V vs. Li/Li⁺) can be at least part-compensated by an outstanding reversible capacity (∼1290 mA h g⁻¹ at 100 mA g⁻¹ after 50 cycles) and a good rate capability (e.g., a capacity of ∼554 mA h g⁻¹ after 20 cycles at 50C). ¹²⁶ Additionally, with respect to energy conversion, the characteristic tuneable direct band gap endows single-layer MoS₂ with the possibility to serve as a viable semiconductor in photovoltaic applications. ¹⁵⁴ In the sensing field, the layered structure and large specific surface area of MoS₂ nanosheets also offer promise for high sensitivity, high-efficiency field effect transistor (FET), optical and electrochemical devices. ¹¹⁷ MoS₂ nanosheets/nanoflakes also exhibit piezoelectric properties, which facilitate their applications in mechanical transducers and piezo-electrocatalysis. ^155,156

MoS₂ is widely feted in various branches of catalysis, for example due to a band gap that couples well with the solar spectrum and on account of the high activity of its edge sites (i.e., the coordinatively-unsaturated sites exposed at the sheet edges). ¹⁵⁷ It has also served as a synergistic catalytic support for gold nanoparticles ¹⁵⁸ and for nickel–iron alloys utilised in the electrooxidation of hydrazine. ¹⁵⁹ Meanwhile, when MoS₂ is grown on the surface of a photocatalyst (e.g., oxygenated monolayer graphitic carbon nitride, O–g-C₃N₄), the chalcogenide's layered S–Mo–S structure is able to suppress agglomeration of the photocatalyst host, which is essential for maintaining the high activity and stability of the photocatalyst. ¹⁶⁰ A further useful feature of exfoliated single-layer MoS₂ is the availability of a myriad of S atoms exposed on the surface of the nanostructure. Sulfur is a soft Lewis base exhibiting a high affinity for heavy metal ions that act as corresponding soft Lewis acids. ^139,147 MoS₂ nanosheets therefore show both a high selective adsorption capacity for certain metal ions, which can be ascribed to abundant sulfur adsorption sites and fast kinetics due to the easy access to these sites. ¹³⁵ Hence, single-layer MoS₂ is among the most effective adsorbent materials for removing heavy metal ions such as mercury from industrial waste water. ¹⁶¹  

Layered bismuth telluride (Bi₂Te₃) crystallises with a rhombohedral structure (space group R3̄m), comprising quintuple layers that are stacked periodically along the c-axis (Figure 1.3(a)). ⁴² As the term suggests, each quintuple layer is composed of five atomic layers that are arranged in the order of –Te(1)–Bi–Te(2)–Bi–Te(1)–, where the numbers in parentheses designate two inequivalent Te sites. Bi atoms are octahedrally coordinated by Te atoms; Te(1) site atoms are bonded with three Bi atoms and three other Te(1) atoms, and Te(2) site atoms are octahedrally coordinated by Bi atoms. ⁴² A complete unit cell is formed by stacking three such quintuple layers in the [0001] direction. ¹⁶² Because the strengths of the ionic and covalent bonds within the quintuple layers are much higher than the van der Waals interactions between them, Bi₂Te₃ is liable to slip along the Te(1) plane during mechanical deformation (i.e., the (0001) acts as a cleavage plane), which is the primary reason for the poor mechanical properties of Bi₂Te₃ single crystals. Bi₂Se₃ and Sb₂Te₃ are isostructural with Bi₂Te₃, and thus the corresponding quintuple layers are –Se(1)–Bi–Se(2)–Bi–Se(1)– and –Te(1)–Sb–Te(2)–Sb–Te(1)–, respectively. These three compounds can form ternary or quaternary continuous solid solutions with tuneable crystal defects, adjustable band gaps and thermoelectric properties that improve upon those of the respective binary compounds. In contrast, Bi₂S₃, Sb₂Se₃ and Sb₂S₃, however, crystallise in an entirely different orthorhombic structure. ^163,164 Nevertheless, rhombohedral Bi₂Te₃, Bi₂Se₃, and Sb₂Te₃, can also form solid solutions with the aforementioned orthorhombic binary chalcogenides over limited compositional ranges without altering their rhombohedral (R3̄m) structure. ^164,165 Beyond these limits rhombohedral and orthorhombic phase separation occurs until a single-phase region of the orthorhombic structure is reached. ^164,165 It is worth mentioning that when more electronegative S or Se substitutes for Te, it is the Te(2) sites that are preferentially occupied compared with Te(1). ^166,167 Such site-selective ordered occupation leads to the minimum energy structure in the system. ^167,168

The band structure of a semiconductor greatly influences its carrier transport properties. ¹⁶⁹ The critical band features, including band shape, band gap, and the number of energy valleys/valley degeneracy (where an energy valley represents the local extreme of the electron energy in the momentum space of the band structure), ^170,171 heavily determine the electrical transport properties of materials. From the band structure of Bi₂Te₃ calculated by density functional theory (DFT) (Figure 1.3(b)), it can be clearly observed that Bi₂Te₃ is a typical indirect band gap semiconductor with multiple valleys, and a narrow band gap of ∼0.13 eV. ¹⁷² The conduction and valence band structure can be represented as a multi-ellipsoidal model with six constant ellipsoids on a reflection plane. ¹⁷² Crystal symmetry causes six degenerate carrier pockets in both the conduction band and valence band of Bi₂Te₃, which benefit the higher density of states and effective mass and in turn contribute to the excellent electrical transport properties. ¹⁶⁹ Similarly, isostructural Sb₂Te₃ and Bi₂Se₃ also have complex, multi-valley band structures. ^42,173

Manipulating point defects and their interplay makes it possible to regulate the carrier concentration and band structure of Bi₂Te₃ and the point defects in Bi₂Te₃ can be categorised into intrinsic and extrinsic types. The former refers to antisite defects and vacancies, while the latter refers to guest atoms introduced generally by doping or alloying. Bi₂Te₃ produced via zone-melting as an ingot is a p-type semiconductor due to antisite defects denoted as Bi′_Te (i.e., Bi atoms occupy the Te positions, providing one hole per antisite defect as Bi has one less electron in the outermost shell than Te). ¹⁷⁴ Halogens are most often used as donor level dopants to tune the electron concentration of n-type Bi₂Te₃, ^175,176 while acceptor levels in p-type Bi₂Te₃ are commonly provided by Ag, ¹⁷⁷ Cu, ¹⁷⁸ Pb, ¹⁷⁹ and Mn dopants. ¹⁸⁰  

Due to its distinctive layered structure and heterogenous bonding features, Bi₂Te₃ inevitably exhibits anisotropic electron/phonon transport. For example, Bi₂Te₃ has both higher electrical and thermal conductivity along the direction(s) perpendicular to the c-axis (in-plane). Previous studies showed that the in-plane electron mobility and hole mobility of Bi₂Te₃ single crystals reached 1200 cm² V⁻¹ s⁻¹ and 510 cm² V⁻¹ s⁻¹, respectively, at 293 K. ¹⁸¹ These figures are accordingly approximately four times and three times those measured for the out-of-plane direction (along the c-axis). ¹⁸¹ Correspondingly, the in-plane electrical conductivity and thermal conductivity are 3–7 times and 2–2.5 times higher than those in the out-of-plane direction. ^182,183 Interestingly, it has been observed that polycrystalline n-type Bi₂Te₃ is easier to texture—i.e., align in the (00l) direction—than its p-type counterpart under the same preparation conditions. ¹⁸⁴ One can hence optimise the anisotropic properties of polycrystalline doped Bi₂Te₃ along a certain direction by regulating the degree of orientation.

It is the small difference in electronegativity between the Bi and Te atoms in Bi₂Te₃ that accounts for the predominantly covalent nature of the Bi–Te bonding. This feature contributes to sharpened electronic bands and in turn the high carrier mobility, ensuring good electrical transport as seen in materials like PbSe. ¹⁸⁵ Meanwhile, it is the high atomic masses of both Bi and Te in Bi₂Te₃, combined with the relatively low melting point of the telluride (∼858 K), that endows phonon vibration with high anharmonicity, thus leading to low lattice thermal conductivity. ^183,186 These qualities combine to make Bi₂Te₃ the most outstanding thermoelectric material currently available for operation near room temperature, and account for its commercialisation in solid-state refrigeration. ^112,187,188 The formation of solid solutions with Sb and Se as substituents engenders a high level of control over the band structure and carrier concentration of Bi₂Te₃, leading to a range of excellent p- and n-type Bi₂Te₃-based thermoelectric materials capable of working across different temperature ranges. ^{60,61,188–190} These solid solutions generally possess better thermoelectric properties and consequently find wider application than the corresponding binary compounds. In addition, Bi₂Te₃ also serves as an example of an important topological insulator with a number of thickness-dependent physicochemical properties. ¹⁹¹ By exploiting its fascinating functional properties, Bi₂Te₃ has already been utilised widely to fabricate various devices for power generation, precision temperature control of microchips, and as a component of self-powered wearable electronics. ^192,193

The development of materials heavily depends on the ability to synthesise them, to explore more cost-effective, energy-efficient synthetic methodologies, and to design novel materials with appropriate compositions and desired functionalities. Synthesis methods make various configurations and microstructures possible for solid materials, which therefore play essential roles in materials innovation. During the past three decades, a variety of synthetic methods have been introduced and implemented in the synthesis of solid-state metal chalcogenides, which include but are not limited to conventional high-temperature melting/heating, hydrothermal (or more broadly, solvothermal) synthesis, vacuum/high-temperature deposition from the gas phase, and emerging techniques such as microwave-assisted synthesis.

Utilising different preparation techniques can generate metal chalcogenides with distinct properties, which depend primarily on chemical composition, crystal structures, microstructure, and crystal defects. The preparation of specific materials requires appropriate regulation and optimisation of synthesis conditions. The following section will discuss the synthetic methodology of metal chalcogenides, which can be divided into three thematic areas: synthesis methods for bulk materials, synthesis of thin films, and nanomaterials synthesis.

The Bridgman–Stockbarger technique is widely utilised for single crystal growth and purification. ^194,195 Raw materials are loaded into a crucible or boat, melted and then resolidified by applying a positive temperature gradient across the melt. This growth process from the melt can be performed in either a vertical or horizontal configuration, ¹⁹⁶ from which the grown crystals generally exhibit either circular or D-shaped cross-sections, respectively. ¹⁹⁷ Crystals with circular cross-section from vertical Bridgman growth are preferred for large-scale epitaxial growth and device fabrication, while horizontal growth encourages high crystal quality due to the reduced thermal stress that results from the configuration. ¹⁹⁷ This method has been applied for the crystal preparation of various metal chalcogenides, such as MSe (M = Ga, Ge, Cd, Sn), ^198–201 MTe (M = Zn, Ga, Cd, Pb), ^{198,199,202,203} SnSe₂, ²⁰⁴ Bi₂X₃ (X = Se, Te), ²⁰⁵ AgSbTe₂, ²⁰⁶ Cu₂SnSe₃, ²⁰⁷ CuInSe₂, ²⁰⁸ Cs₂Hg₆S₇, ²⁰⁹ Cs₂Cd₃Te₄, ²⁰⁹ TlMSe₂ (M = Bi, Sb), ²¹⁰ Tl₂GaInSe₄  ²¹¹ and AgGaGe₃Se₈. ²¹²  

The Czochralski technique has been also used for the preparation of metal chalcogenide single crystals. ^7,197 The apparatus consists of three main components: a crucible, a heater, and a pulling rod positioned axially above the crucible. The growth process starts with melting the required material in a crucible. A seed crystal attached to the end of the pulling rod is then dipped into the melt and slowly lifted while being rotated simultaneously. ¹⁹⁷ The melt temperature and pulling speed are regulated to yield a small meniscus at the end of the seed crystal. This is performed in order to eliminate dislocations and reduce the effects of thermal asymmetry on crystal growth. ¹⁹⁷ The main advantage of this method is the absence of a container for the crystal and so there is no contact between the rod on which the crystal is grown and the melt-containing crucible during cooling. This arrangement eliminates the thermal stress that is normally generated by the mismatch in the thermal expansion coefficients of the container and the crystal, which is important in the production of dislocation-free crystals. ¹⁹⁷ Numerous metal chalcogenide single crystals such as MTe (M = Ge, Sn, Pb), ^213,214 M₂Se₃ (M = Sb, Bi) ²¹⁵ and M₂Te₃ (M = Sb, Bi) ^{190,215–217} have been successfully prepared using the Czochralski technique.

Perhaps the most common way to synthesise polycrystalline metal chalcogenides is by traditional high-temperature solid-state methods, which involves the direct reaction of stoichiometric quantities of high-purity metal and chalcogen element precursors in a sealed ampoule (typically quartz) at relatively high temperatures (ca. 400–1000 °C) in a furnace. ¹ Briefly, certain proportions of reactants are weighed and ground together, pressed into a pellet and then heated to high temperatures to facilitate interdiffusion and compound formation. Considering the propensity of chalcogenides to convert to their corresponding oxides at high temperatures, even when only trace amounts of oxygen are present, reactions are preferentially conducted in either a vacuum-sealed container or an inert-gas stream. After reaction for hours or days, the ampoule is either quenched or slow-cooled to room temperature. Notably, covering reaction crucibles are beneficial to prevent the evaporation of highly volatile components (such as the chalcogen), thus avoiding significant deviation in the composition of the anticipated product materials.

The method typically yields powders that can initially be multiphasic. Post-annealing of the products for several days under vacuum or inert atmosphere without melting is usually required to obtain higher-quality crystals and/or single-phase thermodynamically stable end-products. Many metal chalcogenides can be prepared using this route, including the binary sulfides, selenides and tellurides of transition and main group metals such as Co, Cd, Hg, Pb, Sn, Sb and Ge, ^1,61,218 as well as a variety of complex ternary, quaternary and higher compounds containing elements from all parts of the periodic table (e.g., AgCuTe, ²¹⁹ Pt₃Pb₂Se₂, ²²⁰ K₂MnGe₃S₈, ²²¹ EuHgSnS₄, ²²² EuHgGeSe₄, ²²² Rb₃Ga₃Ge₇S₂₀  ²²³ ).

Chemical vapour deposition (CVD) is a process in which desired thin films or nanostructures are deposited on solid substrates by a heterogeneous chemical reaction of vapour-phase precursors. ²²⁴ In CVD, multispecies of precursors are transported in the gas phase either by evaporation or sublimation through heating, plasma excitation or optical radiation and introduced into a reaction chamber. For example, the synthesis of phase-pure iron pyrite (FeS₂) thin films use di-tert-butyl disulfide and FeCl₃ as the sulfur and iron precursors, respectively, ²²⁵ and the syntheses of MoX₂ (X = S, Se) use Mo(CO)₆ and X (S or Se) as the sources of the respective elements. ²²⁶ When the gas flows over a heated substrate, surface chemical reactions occur to form a desired thin film or nanostructures on a particular substrate, during which further decomposition of the precursors takes place. Defects present in the substrate can act as nucleation sites from which further growth can occur to form desired materials over time. ¹⁹⁷ Surface diffusion and nucleation affect the morphology of the prepared films that can be epitaxial, amorphous or polycrystalline. ¹⁹⁷ Additionally, several factors, such as the deposition temperature, substrate, gas pressure and gas flow rate can strongly affect the crystallinity and morphology of materials prepared by CVD. ^197,227,228 CVD offers a variety of advantages in the deposition of ordered nanostructured thin films with varied compositions, and the technique is suitable for both batch and semicontinuous operations. A series of metal chalcogenide films have been successfully fabricated, ranging from films of binary and ternary compounds such as PbX (X = S, Se), ²²⁹ CuX (X = S, Se), ²³⁰ SnX (X = Se, Te), ²³¹ MTe (M = Ge, Cd), ^232,233 MoX₂ (X = S, Se), ²³⁴ WX₂ (X = S, Se), ²³⁴ ReSe₂, ²³⁵ M₂X (M = Cu, Ag; X = Se, Te), ²³¹ M₂S₃ (M = Cr, Bi), ^236,237 Bi₂Se₃, ²³⁷ M₂Te₃ (M = In, Sb), ^233,238 CdCr₂S₄  ²³⁹ and In₃SbTe₂  ²⁴⁰ films, to uniquely structured films such as Fe-intercalated ZrSe₂, ²⁴¹ NbS₂/MoS₂ heterostructures ²⁴² and ReS₂/WS₂  ²⁴³ heterostructured films.

Physical vapour deposition (PVD) is a commonly utilised technique for the preparation of metal chalcogenide thin films. ¹⁹⁷ In contrast to chemical vapour deposition, PVD realises the synthesis of thin films without chemical transitions from precursors to products. In a typical PVD procedure, target materials with desired compositions are evaporated by various physical means such as by the application of a laser, magnetron or electron-beam under vacuum. ⁸⁶ The generated vapour is deposited onto a heated substrate to form the desired film. ⁸⁶ The substrate is heated to an appropriate temperature and continuously rotated as needed to achieve improved growth homogeneity. ⁸⁶ The basic PVD process is generally divided into three categories: evaporation, sputtering and ion plating in terms of the physical mechanisms of deposition. ^86,197,244 PVD has been applied in the film preparation of a host of chalcogenides such as SnS, ²⁴⁵ Ag₂S, ²⁴⁶ Sb₂X₃ (X = S, Te), ^247,248 M₂Se₃ (M = Bi, As), ^247,249 Ba-doped Ag₂S (Ba:Ag₂S), ²⁴⁶ BaS-doped Sb₂S₃ (BaS:Sb₂S₃) ²⁵⁰ and SnS-doped PbS (SnS:PbS) ²⁵¹ among others.

Chemistry plays a major role in developing new nanomaterials with novel properties of technological importance. ²⁵² Chemical fabrication routes are generally bottom-up strategies that allow a better control of the morphology and dimensions of synthesised nano/microstructures, since various parameters can be regulated individually or simultaneously during preparation. The primary advantage that chemical methods provide is good homogeneity of samples, because chemical synthesis offers mixing and reaction at the molecular level. Methods can be designed to prepare new materials/crystal structures by understanding how matter is assembled and constructed on an atomic and molecular level and the consequent effects on the material's properties. ¹⁰⁶ In this section, several typical chemical synthesis methods of metal chalcogenides are discussed, while other methods for chalcogenide nanomaterials synthesis, such as mechanical alloying and melt spinning, are also introduced.

Hydrothermal and solvothermal methods have been applied for the synthesis of various metal chalcogenide nano/microstructures with specific sizes and shapes. ^{102,253–259} Hydrothermal synthesis involves chemical reactions in an aqueous medium within a closed vessel at temperatures above 373 K and under elevated autogenous pressure (˃0.1 MPa). More broadly, solvothermal synthesis utilises alternative non-aqueous solvents in a similar process. Since such non-aqueous solvents can possess boiling temperatures rather different from water, synthesis can demand modifications to the heating temperature, in turn altering the internal pressure, which provides additional variables towards the successful synthesis of various metal chalcogenides. ²⁶⁰ In addition, solvent properties, such as polarity and viscosity directly impact the solubility, reactivity, and diffusion behaviour of precursors in liquid-phase synthesis, which will further tune the phase, shape, and size of products. ²⁶¹ Various organic solvents, with very different physiochemical properties, are often utilised as reaction media to replace water. Following the successes of hydrothermal and solvothermal syntheses, a promising derivative approach, known as the mixed solvent method, has evolved, which utilises at least two solvents together. The properties of mixed solvents can be regulated by adjusting the components and volume ratios of solvents to achieve the most appropriate synthetic conditions to obtain certain desired products. ^102,258

Solvothermal synthesis methods typically involve the utilisation of sealable reactors composed of a Teflon™-lined vessel and a stainless-steel autoclave. Inert Teflon™ liners are used to avoid any reaction between the reactants and the reaction vessels and to protect the stainless-steel outer shell of the autoclave from corrosive reagents. Figure 1.4(a) provides a schematic diagram of an autoclave and the solvothermal preparation processes. Specifically, chemical precursors in the appropriate ratios are added to pre-added solvent—serving as the reaction medium—in a Teflon™ liner. The Teflon™ liner is sealed in a stainless-steel autoclave, which is heated to a prescribed temperature above the boiling point of the solvent for a certain duration. In the typical solvothermal syntheses of metal chalcogenides, reaction temperatures are usually in the range of 373–523 K and chalcogen sources can include reagents such as S, Se, Te, Na₂S, Na₂SeO₃, Na₂TeO₃, thiourea, SeO₂, TeO₂ and others. ¹⁰⁸ During solvothermal synthesis, pressure is generated in the closed vessel due to solvent vaporisation at elevated temperatures and this increases rapidly with rising temperature. ²⁶² The generated pressure within the reactors also strongly relies on the percentage fill, which is defined in terms of the solvent volume and container volume. The autogenous pressure will rise more rapidly to a higher value when the percentage fill is large. ²⁶³ One should be cautious that if the autogenous pressure is higher than the maximum pressure tolerated by the autoclave, it is possible that the autoclave could explode, causing safety issues. As such, the percentage fill in solvothermal synthesis reported previously is generally not higher than 80%. ²⁶³ It should be also mentioned that most autoclaves have a bursting disc/pressure release valve to prevent such explosions. The starting materials in solvothermal syntheses should be selected with sufficient solubility in the utilised solvent at elevated temperature, so that reactions can continue until completion. The crystallisation of the desired (insoluble) products then occurs at elevated pressure and temperature. It should be noted that variations of solvothermal syntheses can be performed in a microwave field (Figure 1.4(b)), which will be discussed in detail in the section on microwave-assisted solution synthesis.

When compared to solid state ceramic techniques, for example, solvothermal, and mixed solvent methods have many advantages, such as fast reaction kinetics, short processing times, low energy consumption and the high crystallinity of synthesised samples. ²⁵⁴ These methods can realise effective control on the composition, crystallinity, particle size, and microstructure of products by simply regulating reaction variables such as reagent type, temperature, pressure, the presence of additives (e.g., mineralisers, surfactants), reaction duration, solvent ratio, and pH. ²⁶⁰ Considering pH as an example, in alkaline solution OH⁻ can realise the disproportionation of Te into [TeO₃]²⁻ and Te²⁻ through 3Te + 6OH⁻ → 2Te²⁻ + [TeO₃]²⁻ + 3H₂O ^264,265 and in turn promote the formation of metal chalcogenide via the reaction between metal ions and Te²⁻. In contrast, acidic conditions would facilitate the formation of metal–halide bonds by ligand replacement, which is vital to achieving the controllable doping of halides into metal chalcogenides. ^266,267 Variations of solvothermal methods have achieved great success in the preparation of metal chalcogenide nano/microstructures, such as FeS₂ nanowires, ²⁶⁸ FeSe₂ nanoflowers, ²⁶⁹ MnX (X = S, Se) nanorods, ^270,271 ZnSe nanobelts, ²⁷² CoTe nanotubes, ²⁷³ Bi₂Te₃ nanotubes ²⁷⁴ (Figure 1.5(a)), PbTe nanowires ²⁷⁵ (Figure 1.5(b)), T-shaped Bi₂Te₃-Te heteronanojunctions ²⁷⁶ (Figure 1.5(c)), SnS₂ hexagonal nanoplates ²⁷⁷ (Figure 1.5(d)), nano-string-clusters of Bi₂Te₃  ²⁷⁸ (Figure 1.5(e)) and hierarchical flowerlike In₃Se₄  ²⁷⁹ (Figure 1.5(f)). In addition, crystalline products are obtained at relatively lower temperature in solvothermal synthesis methods as compared to melting techniques, for example. However, it should be noted that solvothermal synthesis also presents some clear disadvantages. For example, it is difficult to study the nucleation and growth processes of products in situ as reactions are performed in closed, non-transparent reactors.

Hot injection is an effective synthetic method for the preparation of metal chalcogenides with good crystallinity and narrow particle size distributions. In a hot injection route, a stock solution containing reactive chalcogen precursors is injected quickly into a pre-heated solution containing metal salts. This induces a high level of supersaturation in the reaction mixtures, which in turn leads to a short burst of nucleation that is followed by subsequent growth by aging at a prescribed temperature. ^280,281 Figure 1.4(c) illustrates the hot injection method schematically. The corresponding nucleation takes place instantaneously at a particular temperature. ²⁸² The kinetics of hot injection is such that there is close to negligible time between injection and precipitation. Mechanistic insights have indicated that the separation of the nucleation and growth steps is critical for the formation of monodisperse nanocrystals. ²⁸³ The diameter and morphology of synthesised metal chalcogenide nanocrystals can be effectively tuned by judiciously adjusting the experimental parameters such as type and ratios of starting reagents, injection temperature, reaction duration and use of a surfactant. ^108,284 For example, inclusion of an appropriate surfactant can reduce the surface tension between two liquids or between a solid and a liquid during hot injection synthesis and can also prevent nanoparticle agglomeration by acting as a capping reagent. ²⁶⁰ Hence, the surfactant enhances colloidal stability so that monodispersed nanoparticles can remain stable in a liquid for a long time. Generally, the polar groups of surfactants are used in a reaction medium to bind to the surface of nanocrystals. ²⁶⁰ However, if the surfactant is not completely removed from nanocrystals during the post-synthesis treatment, the residual surfactant may have an adverse effect on the physiochemical properties (e.g., electrical conductivity) of the products. ^285,286

The hot injection method was first used by Murray et al. to prepare monodisperse CdSe nanocrystals in 1993. ⁹⁴ Since then, many other metal chalcogenide nanocrystals have been successfully synthesised via this method, such as CdSe ⁹⁴ (Figure 1.6(a)), Cu₃SbSe₄  ²⁸⁷ (Figure 1.6(b)), ZnS ²⁸⁸ (Figure 1.6(c)), SnSe ^289,290 (Figure 1.6(d) and (e)), Cu₅FeS₄  ⁵⁶ (Figure 1.6(f)), CuS, ²⁹¹ Cu₂S, ²⁹² In₂S₃, ²⁹³ SnTe, ²⁹⁴ GeTe, ²⁹⁵ Ag₂X, ²⁹⁶ PbX, ^297,298 Bi₂X₃, ^299,300 CuInX₂, ^301–303 CuGaX₂ (X = S, Se, Te), ³⁰⁴ SnSe/rGO nanocomposite (rGO = reduced graphene oxide), ³⁰⁵ Cu₂SnS₃, ³⁰⁶ Cu₂ZnSnSe₄, ³⁰⁷ Cu₂CdSnSe₄  ³⁰⁸ and Cu₂ZnGeSe₄  ³⁰⁹ nanocrystals. The particle size and morphology of the metal chalcogenide nanostructures prepared this way are diverse, ranging from nanoparticles and nanorods to nanoplates and nanoflowers. The dominant form depends on the crystal structures of the chalcogenides as well as the synthetic conditions, such as reaction temperature, reaction duration and types of precursors/solvents/surfactants. ^{80,289,310–313}

Physical properties can also be modified by variations in the hot injection protocol. For example, Han et al. could regulate the type of charge carrier in SnSe by changing the Sn precursor and surfactant using an aqueous hot injection method. ^289,290 Following this demonstration, p-type SnSe/rGO nanocomposites were also successfully fabricated through a facile in situ hot injection synthesis method. ³⁰⁵ From these principles, a general surfactant-free, hot injection method was subsequently developed for synthesising binary silver chalcogenides Ag₂X (X = S, Se, Te). ²⁹⁶ Among other examples, Ithurria and Dubertret pioneered the synthesis of ultrathin CdSe nanoplates via hot injection and achieved a precise control of the geometrical parameters of the product crystals (i.e., thickness, crystal shape, lateral size), ³¹⁴ while Ibáñez et al. exerted morphological control in their synthesis of Cu₂CdSnSe₄ to obtain either tetrahedral or penta-tetrahedral shapes. ³⁰⁸ Allen and Bawendi have also reported the controllable hot injection synthesis of Cu–In–Se quantum dots of variable composition and size. ³⁰³ The nanocrystals exhibit red to near-infrared photoluminescence with a direct dependence on both the composition and size of the dots. ³⁰³ The above are just a few examples that demonstrate how the hot injection method can be applied across multiple systems at multiple scales while also providing significant synthetic flexibility in the preparation of metal chalcogenide nano/microstructures.

The utilisation of microwaves in liquid-phase inorganic synthesis was first explored by Komarneni and Roy in 1985. ³¹⁵ Microwaves are electromagnetic waves of frequencies between 0.3–300 GHz, flanked by the infrared and radiofrequency regions of the electromagnetic spectrum. To avoid interference with cable communication and mobile phone frequencies, microwave reactors for chemical synthesis predominantly operate at a specific regulated frequency of 2.45 GHz. ^260,316 This equates to an energy that is insufficient in itself to break typical chemical bonds but is capable of inducing molecular rotations. ²⁶⁰ After decades of development, microwave-assisted synthesis has been widely applied in the fabrication of metal chalcogenides, which is considered as a green synthesis technique. Figure 1.4(b) illustrates the microwave-assisted solvothermal method schematically. Microwaves are non-ionising and only act so as to activate a materials system thermally. ²⁶⁰ Notably, microwave radiation is directly applied to reactants not to reaction vessels. By design, the vessel is effectively transparent to microwave energy, which is absorbed by the reactants and solvents, resulting in efficient heat transfer, principally through dielectric heating. The heating effect of microwaves is triggered by dipolar polarisation, ionic conduction or electrical conduction mechanisms depending on the materials used for reaction. ^260,316 Microwave heating via electromagnetic waves can result in faster reaction kinetics and shorter reaction times (often by orders of magnitude) compared to conventional heating.

The intensity of microwave heating in a reaction mainly depends on the physical properties of the substance placed in the microwave field. ²⁶⁰ The ability of a specific substance to convert microwave energy into heat is determined by the so-called loss tangent (tan δ). ²⁶⁰ A high tan δ material usually has a high dielectric loss, which represents the microwave energy converted to heat. ^260,261 In fact, the microwave-assisted synthesis of bulk chalcogenides can be performed both in the liquid phase and the solid state, as long as the solvent, liquid or solid reactants possess high tan δ values. Microwave-assisted solution synthesis, however, is more appropriate for the syntheses of chalcogenide nanomaterials specifically. Solvents that possess high tan δ values typically also demonstrate good microwave absorption capacity and thus are favoured in the exploitation of the thermal effects achievable using microwave radiation. ^260,261,317 Consequently, polar solvents such as water and alcohol are considered as good media for the microwave-assisted syntheses of inorganic nanostructures in solution. With a high boiling point (471 K), a chemically reducing capability, and a high value of tan δ, ethylene glycol finds wide application in the microwave-assisted syntheses of metal chalcogenides. ²⁶⁰ Generally, therefore, solvents containing dipolar species couple most effectively with microwaves, leading to a rapid rise in temperature and reaction rate.

Microwave heating offers the possibility of rapid chemical synthesis in solution in a very short duration compared to conventional heating, resulting in the potential advantages of lower cost, higher yields, side reaction suppression, and energy saving for chalcogenide nanomaterials production. ^108,260,261 Microwave-assisted solvothermal synthesis employs autoclaves made of high-strength polymeric materials that are transparent to microwaves. ³¹⁸ Apart from the incident microwave power, other reaction parameters, such as the reaction temperature, reaction duration, solvent type and pH also affect microwave-assisted solution synthesis. Through rational design of these parameters, metal chalcogenides with desired chemical composition, diameter, and morphology can be effectively synthesised. For example, Mehta et al. developed a rapid and scalable microwave-assisted solvothermal synthesis of sulfurised Sb₂Se₃ nanowires and nanotubes. ³¹⁹ By adjusting the microwave dose, which represents the product of microwave power and time, the diameter and length of these 1D nanostructures can be controlled. Panda et al. also reported a similar morphology control for metal chalcogenide nanostructures in the microwave synthesis: ³²⁰ by simply tuning the microwave radiation times, the stepwise formation of the ZnS and ZnSe nanowires from small spherical or elongated particles through short rods to long nanowires was observed. Up to now, a variety of metal chalcogenides have been successfully prepared by microwave-assisted solution methods, including binary materials such as HgS, ³²¹ MoSe₂, ³²² SnS₂, ³²³ CdX, ^324–327 ZnX, ^327–330 SnX, ^331–333 PbX, ^327,334,335 CuX, ^336–339 Cu₂X, ^340–342 Ag₂X, ^343–345 Sb₂X₃, ^346–348 and Bi₂X₃  ^349–351 (X = S, Se, Te), as well as ternary and quaternary metal chalcogenides such as AglnS₂, ³⁵² AgIn₅S₈, ³⁵³ CulnX₂ (X = S, Se), ³⁵⁴ CuIn₂S₄, ³⁵⁵ Cu₃SbSe₄  ³⁵⁶ and Cu₂ZnSnS₄. ³⁵⁷  

Liquid exfoliation is regarded as a highly scalable and achievable solution-based method for producing large quantities of thin-layer metal chalcogenides with good crystallinity using mild conditions. ^358–360 This method utilises high shearing or ultrasonication of crystal layers dispersed in a stabilising solvent (such as polymer solutions, ionic liquids, etc.) to produce nanosheets. The thin flakes are then separated from the solvent by centrifugation. Further, two-dimensional metal chalcogenides can be intercalated between layers by various guest species ³⁶¹ and subsequently exfoliated into mono- and few-layer nanosheets by simple ultrasonication. ^{135,362–365} A typical chemical liquid exfoliation procedure involves submerging bulk powders in a solution of an oxidisable Li-containing compound (e.g., n-butyl lithium) for a specific duration to allow the Li ions to intercalate. If this procedure is followed by exposing the intercalated materials to water, then the water will react with the Li between layers, resulting in H₂ generation and in turn the separation of the layers. ^135,363

The type of solvent used in liquid exfoliation directly affects the layer number, thickness, yield, stability and defect chemistry of the exfoliated material. ²⁴⁴ To inhibit the reaggregation of exfoliated materials, the selected solvent must have a surface tension comparable with the surface energy of the layered materials. ²⁴⁴ Coleman et al. reported a universally effective liquid exfoliation method to produce various nanosheets by the dispersion and ultrasonication of a series of layered metal chalcogenides in dozens of solvents with varying surface tensions. ³⁶⁰ Studies of the corresponding correlations were performed using optical absorption spectroscopy to measure the quality of the dispersed materials. Based on the application of theory to the analysis, they proposed that ideal solvents are those that minimise the energy of exfoliation. ³⁶⁰ For example, N-methyl-pyrrolidone and isopropanol emerge as promising solvents for exfoliating numerous layered compounds. Through this method, layered compounds, such as NbSe₂, ^360,366,367 NiTe₂, ³⁶⁰ TiS₂, ³⁶⁸ MoX₂ (X = S, Se, Te), ^{360,366,367,369} WX₂ (X = S, Se), ^{360,366,367,369} TaX₂ (X = S, Se), ^360,367 M₂Se₃ (M = Bi, In) ³⁶⁶ and M₂Te₃ (M = Sb, Bi), ^360,366 can be successfully dispersed, exfoliated and deposited as individual flakes or films.

Mechanical alloying, which is a process of alloy formation in the solid state through the ball milling of high-purity elemental components, has proven to be an efficient method for the preparation of metal chalcogenide nanoparticles. ^370–373 In a typical process, a certain proportion of the respective reactant elemental pieces or powders are loaded into a jar/vessel made of cemented carbides (such as WC), steel or agate, together with a specific number of grinding balls (also manufactured from the equivalent jar material). The tightly sealed jars containing the elemental mixtures are rotated at a predefined frequency, so that the grinding balls experience repeatable mechanical forces resulting from collision, friction, compression, shearing, and grinding. ^370,374 These processes impart mechanical energy to the powders, which makes the powders undergo continuous fracture and welding. The high-intensity, long-duration effects of ball milling ensure that the products are homogeneous, usually resulting in evenly dispersed, ultrafine alloyed particles. ³⁷⁰  

Factors directly affecting the degree of alloying include the type of ball mill, the composition of the grinding balls and the milling conditions employed (e.g., time, speed, ball–powder ratio). ³⁷⁵ The types of instruments for mechanical alloying can vary in the way that energy is imparted and include vibrating, planetary, rotational and high-energy (vibrating–rotational) machines. ²⁶³ It is also possible to mill cryogenically and under different gaseous atmospheres. Tungsten carbide, agate, zirconia, and hardened stainless steel are usually utilised for the construction of jars and the grinding media. The combination of high-energy ball milling and hardened-steel grinding balls generally provides optimal mechanical alloying strength. A reasonable selection of technological parameters, such as the materials of manufacture, the size/dimension of the vessels and the ratio of grinding balls, can improve the ball milling efficiency. ³⁷⁰ As the materials used to manufacture the milling jars and media directly determine key characteristics such as density and hardness, it is necessary to choose the appropriate jars and balls in accordance with the intrinsic properties of the samples to be milled and the desired features of products. The matching of grinding ball material to the jars/vessel material helps avoid erosion and the introduction of unintended impurities into milled mixtures. Additionally, the ball to powder weight ratio is also crucial. ³⁷⁰ Usually, the ball-milling energy increases with the increase of ball to powder weight ratio. However, when the weight ratio exceeds a certain threshold, the limited acceleration space will reduce the collision energy. ³⁷⁰ Ball milling duration also has a significant impact on the phase and microstructure of products; therefore, selecting an appropriate ball milling time is critical for highly efficient preparation and controlled energy consumption. By combining mechanical alloying with various phase and microstructure characterisation techniques, the optimal milling parameters can be tuned to specific metal chalcogenide systems.

Mechanical alloying allows for solid-state reactions without high treatment temperatures. The structural defects of milled products, which are generated by the strong mechanical collisions between balls and between balls and the walls of the container, present plentiful possibilities to obtain various metastable and defect-rich materials with functionalities modulated by these features. ^100,370 For example, p-type nanocrystalline BiSbTe bulk fabricated as a thermoelectric through hot pressing the ball-milled nanopowders, achieved a peak dimensionless figure of merit of 1.4 at 373 K owing to the low thermal conductivity caused by the increased phonon scattering by grain boundaries and defects. ¹⁰⁰ Given the propensity of the method to shift away from equilibrium conditions, milling can enable the generation of alloys that are difficult to obtain via solidification processes due to unfavourable thermodynamics (as depicted by phase diagrams). ³⁷³ A tremendous variety of metal chalcogenide nanomaterials, including stoichiometric compounds such as PbTe, ³⁷⁶ SnS₂, ³⁷⁷ SnX, ^378–380 Ag₂X, ³⁸¹ AgCuTe, ³⁸² CuInX₂, ²²⁴ CuCrS₂  ³⁸³ and CuCr₂X₄ (X = S, Se, Te), ³⁸⁴ as well as solid solutions and composite architectures such as Sn(S,Se)₂, ³⁷⁷ (Bi,Sb)₂Te₃  ¹⁰⁰ and CNTs/Cu₂Se ³⁸⁵ have been prepared this way.

Solidification is a conventional method to fabricate metals and alloys, which involves heating a material above its melting point and then allowing the melt to cool to room temperature at various speeds. ³⁸⁶ The heating sources for melting are versatile, such as electrical resistance heating, induction heating, arc melting, and levitation melting. ^387,388 In particular, the melt spinning technique, involving the rapid cooling of melts, has been widely applied to prepare amorphous alloys. ^389,390 Figure 1.7(a) presents a schematic illustration of melt spinning, revealing that in a typical process, a thin stream of melt is ejected onto a high-speed rotating Cu wheel (that is cooled internally, usually by water) under a protective Ar atmosphere. Due to the ultra-high cooling rate of 10⁴–10⁷ K s⁻¹, ³⁹¹ the melt contacting the Cu wheel will cool relatively quickly and form a nanostructured ribbon-like material with a relatively uniform grain size distribution or even an amorphous material. ^386,392 This method has been applied to prepare a variety of nanostructured metal chalcogenides, such as Bi₂Te₃–, ^393–395 AgSbTe₂–, ^396,397 SnTe–, ³⁹⁸ and PbTe–based materials. ³⁹⁹  

The microstructures generated in melt spun materials are sensitive to their positions relative to the rotating Cu wheel. ^398,400,401 In general, melt spinning will create ribbons with two entirely different surfaces: the surface that contacts the Cu wheel is defined as the contact surface, while the opposing one is called the free surface. These two surfaces have significantly distinct cooling rates, temperatures, critical radii of nucleation and growth rates of nucleation sites, leading to the formation of multiscale microstructures. ^391,400 Progressing from the contact surface towards the free surface of the ribbons, one can observe an evolution from amorphous microstructures, through fine nanocrystalline regions, to larger nano/micro-sized crystals. ^263,400 This is what occurs in Bi₂Te₃-based materials for example (Figure 1.7(b) and (c)) and the contact surface of a Bi_0.48Sb_1.52Te₃ ribbon features an amorphous structure embedded with 5–10 nm nanocrystalline domains, while the free surface is characterised by micro-sized dendritic crystals with nanoscale features. ⁴⁰⁰ Significantly, the nanostructures of melt-spun ribbons are determined not only by the melt spinning process but also by the properties of the starting materials, so different materials could possess totally different nano/microstructures. ^398,402 For example, Yan et al. controllably synthesised SnTe ribbons consisting of columnar grains via melt spinning (Figure 1.7(d) and (e)) and found that Sb doping effectively hindered crystal growth during solidification, leading to a decrease in the diameters of the SnTe-based columnar grains. ³⁹⁸ To summarise, adjusting the chemical compositions of the raw materials and the cooling conditions during melt spinning can effectively dictate the growth of different nanostructured chalcogenide materials.

Energy serves as the backbone for modern society. The accomplishments of civilisation have mainly been achieved through the increasingly extensive and efficient harnessing of multi-form energy to extend human ingenuity and capability. The world's routine consumption of energy is confronted by a series of serious challenges in the context of long-term sustainability, which include the depletion of traditional non-renewable fossil fuel reserves, rapidly surging energy demands, global warming, and other energy-relevant environmental problems. ^403,404 As a sustainable energy supply is essential for global development, the exploration of novel high-efficiency strategies for clean, renewable energy storage and conversion is becoming increasingly urgent. ^404,405 Among different types of materials, metal chalcogenides, exhibiting rich chemical compositions, diverse crystal structures, tuneable nano/microstructures, and fascinating physicochemical properties, have been considered as promising building blocks for renewable energy conversion and storage devices. ^{6,72,108,406–408} In addition, many metal chalcogenides possess several other advantages with respect to the selection of materials for devices, such as relatively low cost and high Earth abundance, each of which are of great significance for practical (commercial) application. ¹⁰⁸ Metal chalcogenides have attracted wide attention in a variety of promising energy-related applications, such as lithium-ion batteries, ^409–412 supercapacitors, ^413–418 thermoelectric conversion, ^{109,188,419–425} solar cells ^{9,58,298,426,427} and catalysis. ^428–432 This section will focus on these typical applications of metal chalcogenides.

Efficient energy storage is a long-standing scientific and technological issue that has global effects on humanity. Electrochemical techniques with the advantages of environmental friendliness and high efficiency are regarded as ones of the most promising candidates for clean energy storage. ^433–435 Secondary batteries, also called rechargeable batteries, comprise cells constructed for storing and releasing electrical energy in which each cell, after discharge, can be restored to its original charged state. Metal chalcogenides feature in various battery chemistries, including lithium-ion, sodium-ion and lithium–sulfur batteries among others. ^436–440  

A lithium-ion battery (LIB) refers to a type of secondary battery, which utilises internal lithium ions and external electrons to move between the positive and negative electrodes to convert chemical energy into electricity and vice versa. LIBs entered the energy stage in the early 1970s and have been quickly developed since their first commercialisation by the Sony Corporation in 1991. ^441,442 The main components of a LIB include a positive electrode (i.e., cathode), a negative electrode (i.e., anode), an electrolyte and a separator. The operation principles of LIBs can be summarised as follows (Figure 1.8): (1) when LIBs are charging, Li⁺ ions de-intercalate under an applied potential from the cathode and then intercalate into (react with) the anode; (2) when discharging, Li⁺ ions transport through the electrolyte to intercalate back into (react with) the cathode, generating a current. ^443,444 During such charge/discharge cycles, Li⁺ ions shuttle back and forth between the cathode and anode, realising the interconversion between chemical energy and electrical energy and enabling electrochemical energy storage within a battery. Accordingly, LIBs are also referred as ‘rocking chair batteries’, as originally proposed by Armand et al. in 1980. ⁴⁴⁵  

LIBs have already broadly served as the predominant power supply for consumer electronics and electric vehicles for several decades, owing to their relatively high energy density, low self-discharge rate and good cycle life. ^446–451 Commercial LIBs generally utilise graphite as anodes and oxide Li-ion intercalation host materials (e.g., LiCoO₂) as cathodes. However, these electrode materials experience several limitations. For example, the relatively low theoretical capacity (372 mA h g⁻¹) and slow Li⁺ diffusion rate (10⁻⁸ cm² s⁻¹) of the graphite anode impose restrictions on the energy density and power density of LIBs, respectively. ¹⁰⁸ Moreover, the growing demands for lighter and thinner LIBs with higher capacity has boosted ongoing research for new electrode materials with superior properties. Therefore, tremendous efforts have been devoted to the exploration and design of alternative electrode materials with high energy storage performance. ^452–455 Among them, metal chalcogenides especially layered ones have been regarded as electrode materials of immense potential for LIBs due to the possibilities of high capacity and long cycling life. ^409,456

Typically, MoS₂ has been widely investigated as an anode material for rechargeable batteries since its layered structure and large specific surface area can provide plentiful active sites for the reversible electrochemical insertion of guest ions. ^457–459 Additionally, MoS₂ exhibits good capacity retention given the high reversibility of redox reactions with Li ions. ⁴³⁶ A high specific capacity of 750 mA h g⁻¹ has been reported when using restacked MoS₂ single layers as electrodes. ⁴⁶⁰ However, MoS₂ has some intrinsic shortcomings: for example, the electrochemical performance is downgraded by the unfavourable reaction kinetics caused by its low electrical conductivity and the relatively large volume changes that occur during the charge and discharge processes. ⁴⁶¹ Fortunately, the above problems can be mitigated via the modification of MoS₂ using a number of different strategies. Introducing defects—defect engineering—has been shown to endow MoS₂ with beneficial characteristics, such as providing additional electrochemically active sites for Li⁺, resulting in optimised electrochemical performance. ^410,412,462 For example, vacancies introduced into MoS₂ break the periodic crystal structure, leading to a much decreased interlayer diffusion energy barrier; ⁴⁶² Qin et al. reported N-doped mesoporous MoS₂ nanosheets with enhanced Li⁺ ion storage performance (a consistent capacity of 998 mA h g⁻¹ after 100 cycles), which is a consequence of their large surface area, porous microstructure and enlarged interlayer distance. ⁴¹⁰  

Additionally, the interplanar distance can play a decisive role in facilitating Li ion intercalation and diffusion in an MoS₂ anode. Liu et al. prepared a highly ordered mesoporous MoS₂ material with an expanded (002) interplanar spacing of 0.66 nm, providing sufficient space for ultra-fast Li⁺ ion intercalation. ⁴⁶³ In fact, the three-dimensional diffusion of Li⁺ is of vital importance to achieve fast charge/discharge kinetics. Tao et al. synthesised defect-enriched few-layer MoS₂ nanosheets for Li⁺ ion storage and determined that the energy barrier for Li⁺ ions to pass through the double-vacancy MoS₂ layer is only 0.238 eV, which is dramatically lower than the zero-defect (3.14 eV) and single-defect (7.97 eV) MoS₂. ⁴¹¹ In addition, MoS₂ can be incorporated with other capacitive and conductive materials to yield MoS₂-based hybrid LIB electrodes with improved electrochemical performance and stability. ^{412,464–466} For example, MoS₂/graphene composites exhibited a high reversible capacity of 1100 mA h g⁻¹ at a current density of 100 mA g⁻¹ and good cycling stability when applied as an anode in an LIB. ⁴⁶⁵ Similarly, MoS₂/CNT hybrids achieved enhanced ionic transfer, thus yielding an extraordinary capacity of 1320 mA h g⁻¹ at a current density of 100 mA g⁻¹ and an impressive cycling stability of up to 1000 cycles without appreciable capacity fade. ⁴⁶⁶  

Apart from Mo₂S, a variety of metal sulfides, selenides, and tellurides have also been explored and developed as high-performance LIB electrode materials. These include: MS (M = Zn, Mn, Fe, Sn, Cu, Ni, Co), ^467–473 MSe (M = Sn, Zn, Fe), ^474–476 SnTe, ⁴⁷⁷ MS₂ (M = Mo, W, Sn, Zr, Ti, Fe, Ni, V, Co), ^478–485 MSe₂ (M = Mo, W, Co, Ni, Fe, Sn), ^{474,475,478,486,487} Cu₂S, ⁴⁸⁸ M₂S₃ (M = In, Sb), ^489,490 M₂X₃ (M = Sb, Bi; X = Se, Te), ^491–495 VS₄, ⁴⁹⁶ In₃Se₄, ⁴⁹⁷ Ni₃S₄  ⁴⁹⁸ and Co₉S₈. ⁴⁹⁹  

The supercapacitor (SC) is considered as one of the most promising energy storage systems due to characteristic features such as high power density, fast charge/discharge, a long cycling lifespan, simple fabrication procedures and relatively low cost. This combination of features yields a device that bridges the gap between conventional capacitors and batteries. ^500–505 Current commercial SCs are mainly based on carbon materials, which normally have limited energy densities (less than 10 W h kg⁻¹). ⁵⁰⁶ A higher energy density is certainly desirable for almost all practical applications of SCs, so intensive efforts have been devoted to developing alternative electrode materials with higher specific capacitance. ^414,507,508 Owing to the amalgamation of several intrinsic physicochemical properties coupled with good electrochemical performance, several metal chalcogenides have been proposed as electrode materials for SCs. ⁴¹⁵ For example, two-dimensional layered architectures comprising metal chalcogenides and other conducting materials (e.g., chalcogenide/graphene) have been regarded as among the most promising electrode materials for high performance SCs. ⁵⁰⁹  

A SC consists of two working electrodes (composed of active materials deposited on current collectors), an electrolyte, and a separator that electrically isolates the two working electrodes. According to the different possible charge storage mechanisms, SCs can be classified into three types: electric double-layer capacitors (EDLCs), pseudo-capacitors and hybrid supercapacitors. ^{414,508,510,511} The function of EDLCs is controlled by the reversible adsorption/desorption of electrolyte ions at the electrode/electrolyte interface, which forms an electrical double layer region. ^512,513 The energy storage achievable through this physical process is relatively low and highly dependent on the accessible surface area of the electrode materials for the adsorption/desorption of ions; high surface area carbon-based materials are typically excellent candidates to serve as EDLC electrodes. ⁵¹⁴ In contrast, pseudo-capacitance is based on a quick and reversible faradaic charge transfer through redox and/or intercalation reactions in active materials. ⁵¹⁵ These processes endow SCs with much higher charge storage capacities than EDLCs. The charge storage made possible via faradaic reactions in metal chalcogenides generates appreciably higher energy density in pseudo-capacitors when compared with carbon-based materials. ⁵¹⁶ Hybrid SCs combine the concepts exploited in both EDLCs and pseudo-capacitors and contain two electrodes with very different characteristics; the electrode on one side demonstrates electrostatic capacitance, while the electrode on the other side functions via pseudo-capacitance. ^413,511,517 Similar to any other energy storage system, the performance of an SC is mainly evaluated in terms of the specific capacity, the energy density and the power density.

Two-dimensional layered MoS₂, which comes with the advantages of an ultrahigh surface area, excellent mechanical properties, high ionic conductivity and high theoretical capacity, is regarded as one of the most outstanding candidates for SCs. ^{418,518–520} For example, Fan et al. successfully prepared flower-like MoS₂/C nanospheres with a high specific capacitance of 201.4 F g⁻¹ at a current density of 0.2 A g⁻¹. ⁵²⁰ Acerce et al. reported that chemically exfoliated MoS₂ nanosheets containing a high proportion of metallic 1T phase can electrochemically intercalate ions such as H⁺, Na⁺ and K⁺, which therefore obtain high capacitances ranging from 400 to 700 F cm⁻³ in different aqueous electrolytes. ⁴¹⁸ The resulting 1T nanostructured MoS₂ film electrode was operated at a high voltage of 3.5 V in organic electrolytes, exhibiting favourable volumetric energy and power densities, as well as a high Coulombic efficiency of 95% after 5000 cycles. The excellent electrochemical performance is ascribed to the hydrophilicity and high electrical conductivity of 1T MoS₂, as well as the ability of the exfoliated layers to dynamically expand and intercalate various ions. ⁴¹⁸ Further, the synergetic effects that exist between MoS₂ and carbon materials can facilitate charge transport and electrolyte diffusion in C/MoS₂ composites and relieve the volume expansion/contraction of the electrode during cycling. ^509,521,522 For example, da Silveira Firmiano et al. reported that an rGO/MoS₂ hybrid electrode achieved an energy density of 63 W h kg⁻¹ and an exceptional cycling stability with 92% specific capacitance remained after 1000 cycles. ⁵⁰⁹ Additionally, the coating of transition metals/metal oxides/metal sulfides on MoS₂ works so as to both improve the structural stability and electrochemical performance of MoS₂. ⁴¹⁷ For example, Ti/TiO₂/MoS₂ coaxial hybrids demonstrated a capacitance of 230.2 F g⁻¹ and an energy density of 2.70 W h kg⁻¹ when adopted in a symmetrical SC, in addition to possessing both mechanical flexibility and stability. ⁴¹⁷  

Alternatively, NiCo₂S₄ with a high redox activity has become very attractive as an SC electrode due to its high electrical conductivity, electrochemical activity, and capacity compared to mono-metallic sulfides. ^416,523 Chen et al. reported that sea-urchin-like NiCo₂S₄ exhibited dramatically improved electrochemical properties when compared to the sulfides of either of the individual metals, Ni or Co. ⁵²⁴ The NiCo₂S₄ electrodes also exhibited high storage stability and long lifetimes with a 91.4% specific capacity retention after 5000 cycles at a current density of 20 A g⁻¹. ⁵²⁴ Later, Pu et al. successfully prepared hollow hexagonal nanoplates of the same chalcogenide, NiCo₂S₄, which displayed a high specific capacitance of 437 F g⁻¹ at a current density of 1 A g⁻¹ using 3 mol L⁻¹ KOH electrolyte solution. ⁵²⁵ Crucially, it was Chen et al. that showed that the electrochemical performance of NiCo₂S₄ was strongly associated with the Ni/Co atomic ratio and that the best performance was found in a composition with an Ni/Co atomic ratio of 1. ⁵²⁶ In addition, Ouyang et al. reported that an S-Co₃O₄@NiCo₂S₄ hierarchical structure exhibited battery-like behaviour with a remarkably large areal capacity of 10.9 mA h cm⁻² at a current density of 8 mA cm⁻², together with an outstanding cycling stability, demonstrating a capacity retention of 97.3% after 5000 cycles. ⁵²³ The excellent electrochemical performance of this hierarchical composite was attributed to plentiful electrochemically active sites, fast electron transport paths and the expanded surface area provided by the binder-free core–shell structure. ⁵²³  

A large proportion of energy generated worldwide is lost as waste heat. Effective recovery of waste heat is crucial for solving energy and environmental problems by providing supplemental electrical power without burning additional fossil fuels. Fortunately, thermoelectric technology offers a reliable and environmentally friendly solution to harvest electrical energy from the recycling of waste heat. ^{109,405,527,528} Additionally, thermoelectric technology can be utilised to convert heat produced by concentrated or unconcentrated sunlight into electricity. ^529,530 This is important because infrared radiation with photon energy below the band gap of photosensitisers is not absorbed in conventional photovoltaic cells and generates only waste heat. ⁵²⁹ The three thermoelectric phenomena have been known since the respective discoveries made by Thomas Johann Seebeck, Jean Charles Athanase Peltier, and William Thomson in the 19th century and enable the direct and reversible conversion between heat and electricity in an all-solid-state system without noise, vibration, or emissions of greenhouse gases. ^531–533 Solid-state power generation and refrigeration systems, based on the Seebeck and Peltier effects, respectively, have received intensive and sustained attention. ^{527,534–536} Figure 1.9(a) and (b) provide schematic diagrams of the Seebeck and Peltier effects, respectively. Thermoelectric devices possess many advantages, such as no moving parts and long lifespans that require no maintenance for long periods. These traits favour their use in fields such as remote power supplies (e.g., radioisotope thermoelectric generators), ^537–539 the automotive industry, ^540,541 temperature sensing, ⁵⁴² temperature control devices, ⁵⁴³ and implantable/wearable devices. ^544–546  

where S, σ, T, κ _e and κ _l are the Seebeck coefficient, electrical conductivity, absolute temperature, electronic thermal conductivity and lattice thermal conductivity, respectively.

where n is the carrier concentration, e is the electron charge, µ is the carrier mobility, k _B is the Boltzmann constant, T is the absolute temperature, h is the Planck constant, m* is the density of states effective mass, and L is the Lorenz number. Note that the expression for S (eqn (1.5)) is valid for a degenerate semiconductor with a parabolic band dispersion. It can be seen that these three parameters (σ, S and κ _e) are coupled strongly through the carrier concentration n, which makes it difficult to optimise one parameter without (adversely) affecting the other parameters. Therefore, it is necessary to explore strategies that can realise either independent or synergistic regulation of electron/phonon transport in order to optimise thermoelectric performance. After long-term development, researchers have learnt to exploit numerous strategies to improve thermoelectric performance. This might be achieved by enhancing S ² σ via energy band engineering, ¹⁷⁰ optimising carrier concentration, ⁵⁵¹ introducing resonant states ^552,553 and/or utilising energy filtering effects. ⁵⁵⁴ Alternatively, performance could be improved via suppressing κ _l through nanostructure engineering, ⁵⁵⁵ defect engineering ^556,557 and/or by forming nanocomposites ^305,385 and constructing multi-scale hierarchical structures, ⁵⁵⁸   etc. The current priority for developing high performance materials in the thermoelectric field remains to optimise the property parameters synergistically and to integrate all the above strategies effectively.

Generally, a good thermoelectric material behaves as a ‘phonon-glass electron-crystal’ (PGEC), with the excellent electrical properties typical of a crystalline solid and the ultralow thermal conductivity typical of a glass. ^559,560 To date a number of thermoelectric materials, including metal chalcogenides, ^{100,108,424,425,561–566} skutterudites, ^567–569 clathrates, ^570–572 Zintl phase compounds, ^573–576 complex alloys, ^422,577,578 conductive polymers, ^579–581 and some oxides, ^582–584 have been identified as promising systems. Among them, metal chalcogenides have already demonstrated outstanding thermoelectric performance and with the huge structural and compositional flexibility that is manifest with chalcogenides, they offer a good platform for designing new and improved materials. ^{181,263,547,560} For example, Bi₂Te₃ and its solid solutions with Se or Sb are archetypal thermoelectric materials for operation in devices near room temperature, ^112,187,193 and PbTe-related materials are prime candidates for medium-temperature applications; ^{419,585–587} both these systems are representative of traditional and well-established thermoelectric materials. In recent years, many new high performance thermoelectric chalcogenide materials have been identified and widely explored. SnSe- and Cu₂Se-based materials are typical of the new focus for thermoelectrics providing not only some of the highest ZT values in existence, but also meeting ecological requirements of Earth-abundant component elements and low environmental impact. ^{385,421,425,588–590} We now focus principally on these four metal chalcogenide systems (Bi₂X₃, PbX, SnX and Cu₂X), their fundamental chemistry and their progress to date as bulk thermoelectrics.

As state-of-the-art near-room-temperature thermoelectric materials, (Bi,Sb)₂(Te,Se)₃ were first commercialised and utilised in thermoelectric refrigerators several decades ago. ⁵⁴⁷ Across all the possible solid solutions, the compounds Bi₂Te_2.7Se_0.3 and Bi_0.5Sb_1.5Te₃ correspond to the optimal compositions of n- and p-type high-performing thermoelectric materials, respectively. ⁵³⁰ Layer-structured Bi₂Te₃ demonstrates anisotropic thermoelectric properties with higher ZT values in the ab plane than along the c-axis. ^182,183 Therefore, one way to enhance the ZT of Bi₂Te₃-based compounds is to influence the orientation of crystals. For example, hot deformation has been proven to be an effective method to increase the degree of texturing (preferred orientation) and to optimise the concentration of defects in polycrystalline Bi₂Te₃. ^591–594 Consequently, the ZT values of the corresponding n-type and p-type Bi₂Te₃-based materials can reach levels of 1.2 (445 K) ⁵⁹¹ and 1.3 (380 K), ⁵⁹⁵ respectively. Nanostructuring can also enhance thermoelectric performance by way of reducing lattice thermal conductivity through introducing high-density grain boundaries. ^{100,112,596–598} For example, Poudel et al. achieved significantly improved ZT values (a peak ZT of 1.4 at 373 K) in a p-type Bi–Sb–Te nanostructured bulk material prepared by ball milling followed by hot pressing. ¹⁰⁰ The magnitude of the figure of merit was mainly ascribed to the low lattice thermal conductivity resulting from the intensified phonon scattering by grain boundaries and defects; ¹⁰⁰ similarly, p-type nanostructured Bi_0.5Sb_1.5Te₃ prepared by combining melt spinning and spark plasma sintering reached a maximum ZT of 1.56 at 300 K. ⁵⁹⁸ The thermoelectric performance of Bi₂Te₃-based alloys can also be drastically improved through the introduction of nanoscale secondary phases into matrices. ⁵⁹⁹ For example, (Bi,Sb)₂Te₃ composites embedded with nanostructured SiC achieved a ZT of 1.33 at 373 K. ⁵⁹⁹ Additionally, anti-site defects and vacancies in Bi₂Te₃ can play decisive roles in the conduction type and carrier concentration. ^{561,597,600,601} Hence, defect engineering can be used very successfully to tune the thermoelectric performance of Bi₂Te₃-based compounds. A further example of this approach was provided by Kim et al. who adopted a liquid-phase compaction method to create dislocation arrays at grain boundaries in Bi_0.5Sb_1.5Te₃. ⁶⁰⁰ In this way they were able to obtain very obvious reductions in lattice thermal conductivity, leading to a maximum ZT of 1.86 at 320 K. ⁶⁰⁰  

Lead chalcogenides, especially PbTe-based materials, are traditional high-performance thermoelectric materials for mid-range temperature (500–900 K) applications. ^{585,586,602,603} The compounds crystallise with the face-centred cubic NaCl crystal structure (space group Fm3̄m). Various dopants have been applied to tune the conduction type and carrier concentration. For example, acceptor doping (p-type) can be achieved by placing Li, Na and K onto the Pb sites, while donor doping (n-type) is possible by either introducing Ga, In, La, Sb, Al, Bi onto the Pb sites or by doping Cl, Br, I on the Te sites. Among these options, Na has proved to be a very effective dopant for p-type PbX (X = S, Se, Te) systems, possessing the highest solubility in PbS (2 mol%) and somewhat lower solubilities in PbSe (0.9 mol%) and PbTe (0.5 mol%). ⁶⁰⁴ Na doping also introduces point defects and precipitates, leading to a reduced lattice thermal conductivity. ⁶⁰⁴ Regarding n-type PbX, the best performing materials have been proven to be Al-doped PbSe ⁶⁰⁵ (with a ZT of 1.3 at 850 K) and I-doped PbTe ⁶⁰⁶ (ZT = 1.4 at 700–850 K). Creating nanoprecipitates and/or building hierarchical structures can additionally significantly reduce the lattice thermal conductivity of PbX. ^{558,602,607–610} Wu et al. reported that 3% Na-doped (PbTe)_0.8(PbS)_0.2 achieved a maximum ZT of 2.3 at 923 K due mainly to the extremely low thermal conductivity arising from the phase boundaries, ⁶¹⁰ whereas Biswas et al. achieved significantly reduced lattice thermal conductivity and dramatically enhanced thermoelectric performance (a ZT of 2.2 at 923 K) in PbTe-based materials by constructing multi-length scale hierarchical architectures that include point defects, nanoscale endotaxial precipitates and mesoscale grain boundaries in the same material. ⁵⁵⁸ Such architectures are capable of scattering phonons across different wavelengths. ⁵⁵⁸ Jiang et al. realised a superior ZT of 1.8 at 900 K in an n-type PbSe-based high-entropy material that consisted of each of Pb, Sb, Sn, Se, Te and S in a chalcogenide of chemical formula of Pb_0.89Sb_0.012Sn_0.1Se_0.5Te_0.25S_0.25. ⁶¹¹ Such a high ZT was attributed to first, the consistent electronic transport properties that result from a stabilised single-phase structure and second, the ultra-low lattice thermal conductivity that arises from the strong phonon scattering across a highly strained lattice. ⁶¹¹  

Tin selenide (SnSe) has a layered orthorhombic structure derived from a three-dimensional distortion of the rock-salt structure and possesses a strong anisotropy. ^425,588 SnSe undergoes a second-order displacive phase transition from the Pnma space group to a higher-symmetry Cmcm space group when the temperature increases above ∼750 K. ⁵⁸⁸ SnSe is one of the most promising Te-free thermoelectric candidates emerging in the recent decade and meets requirements of environmental friendliness and low cost in addition to outstanding thermoelectric performance. ^425,589,612 Specifically, the low lattice thermal conductivity is due to a strong anharmonicity, and the excellent electrical transport properties result from multiple electronic valence bands. ^425,551,589 SnSe-based materials have garnered great attention and experienced vigorous development over a short period of time. ^420,613 For example, high ZT values of ∼2.6 at 923 K  ⁴²⁵ and ∼2.8 at 773 K  ⁶¹² were reported in p-type Na-doped SnSe and n-type Br-doped SnSe single crystals, respectively, which originate from both a very low lattice thermal conductivity and the high magnitude of the power factor. Meanwhile, the relatively less impressive thermoelectric performance of polycrystalline SnSe has also been improved by various strategies, such as doping, ^{378,614–616} forming solid solutions with other chalcogens, ^80,617–619 nanostructuring, ^614,620,621 obtaining higher purity samples with minimal oxidation ^331,622 and composite engineering. ^{305,340,623–626}

p-Type Cu₂Se-based materials constitute the other main group of candidate thermoelectrics that have been explored extensively in the past decade and their popularity originates from their phonon-liquid electron-crystal (PLEC) behaviour. ^{385,423,424,590,627} Cu₂Se has a unusual cubic structure at elevated temperature, where ordered Se anions form a rigid sublattice and provide pathways for charge carriers, while Cu cations are randomly distributed around the static Se. ⁴²⁴ The mobile Cu ions in Cu₂Se can scatter phonons significantly and eliminate partial transverse vibrational modes, which leads to extremely low lattice thermal conductivity while preserving good electrical conductivity. ⁴²⁴ Liu et al. first reported this liquid-like Cu⁺ transport behaviour in non-stoichiometric Cu_2−x Se and realised a high ZT of 1.5 at 1000 K inspiring systematic investigations of the selenide's PLEC phenomenon and encouraging multiple efforts to enhance thermoelectric performance further. ⁴²⁴ Ball milling and hot pressing was found to induce an extremely low lattice thermal conductivity (0.4–0.5 W m⁻¹ K⁻¹) and a ZT of ∼1.6 at 973 K; ⁵⁵⁹ as with other chalcogenide thermoelectrics, substitution and nanostructuring strategies also proved extremely successful. ⁶²⁸ Cu_1.94Se_0.5S_0.5 solid solutions achieved a remarkable ZT of 2.3 at 1000 K, ⁶²⁸ whereas formation of a Cu₂Se/CuInSe₂ nanocomposite doped with 1 mol% In improved chemical stability of the selenide and yielded an extraordinary peak ZT value of 2.6 at 850 K. ⁶²⁷ At a likely different length scale, Cu₂Se/CNT hybrids achieved a remarkable high ZT of 2.4 at 1000 K due to the homogeneously dispersed CNTs inside the Cu₂Se matrix and correspondingly formed Cu₂Se/CNT hybrid interfaces. ³⁸⁵ Apart from Cu₂Se, other Cu-based PLECs, such as Cu₂S, ⁶²⁹ Cu₅FeS₄, ^{55,56,630,631} and CuAgSe, ⁶³² have also attained excellent ZT due to low lattice thermal conductivity and improved electrical transport properties.

In addition to chalcogenides of Bi, Pb, Sn and Cu, Ag₂X (X = S, Se, Te), ^{296,633–637} GeX (X = Se, Te) ^638,639 and various indium selenides (e.g., In₄Se₃, In₃Se₄) ^7,562 have also been explored for thermoelectric applications. Orthorhombic-structured Ag₂Se, for example, demonstrates high ZT near room-temperature due to its large carrier mobility and low lattice thermal conductivity. ^296,636,637 The former is related to the low band effective mass and the latter is ascribed to the low cut-off frequency and group velocity as well as the hybridisation of the optical and acoustic phonon branches. ⁶³⁶ Recently, Ag₂Se/CNT composites were synthesised by in situ solution synthesis and spark plasma sintering; due to the interfacial phonon scattering and energy filtering effect, the optimal Ag₂Se/CNT sample registers a remarkably low κ _l of 0.27 W m⁻¹ K⁻¹ and a slightly enhanced power factor of 2.71 mW m⁻¹ K⁻² at 375 K, leading to an exceptional peak ZT of 0.97 at 375 K. ⁶³⁶ Similarly, the AgSbSe₂ component in Ag₂Se/AgSbSe₂ composites enables intensified phonon scattering and impeded crack propagation, leading to an excellent average ZT of ∼0.89 between 300 and 375 K and a compressive strength of 197.3 MPa. ⁶³⁷ Undoubtedly, metal chalcogenides have established themselves as most promising thermoelectrics, particularly in the low and medium temperature ranges.

Sunlight is the most widely and abundantly distributed energy source and the solar energy that strikes the Earth in a day is much more than the total energy consumption of the world in a year. ^640,641 Energy harvesting directly from sunlight provides a desirable solution for fulfilling the needs of clean energy without negative environmental impact. Solar cells or photovoltaics can capture sunlight and convert solar energy into electricity. These are therefore excellent devices for solar power generation and have already been commercialised, with most of the solar cells that have been utilised based on crystalline silicon. ^427,642 High-purity silicon is expensive and so during the past few decades substantial efforts have been devoted to exploring and tailoring various semiconductors for solar cells with lower manufacturing costs compared to silicon without incurring a significant loss of conversion efficiency. ^643–647 Metal chalcogenide semiconductors such as cadmium telluride (CdTe) and copper indium gallium selenide, Cu(In,Ga)(Se,S)₂, possess outstanding photovoltaic performance but can be produced relatively cheaply compared to silicon. ^426,648 The chalcogenide materials have therefore been considered as key materials in the development of high-efficiency solar cells. ^649,650

A solar cell is a semiconductor-based device that serves to convert light energy (certain wavelengths) into electricity via the photovoltaic effect. The key indicator in assessing the performance of a solar cell is the amount of electrical power that can be extracted from the light available to it. The size (energy) of the band gap of the semiconducting photoelectrode is inherently important since it should be suitable for the utilisation of the maximal portion of the solar radiation spectrum. ⁶⁵¹ The thickness of the semiconducting photoelectrode needs to be large enough to absorb the maximum number of photons of incident solar radiation. The working principle of energy conversion in a solar cell involves three basic processes, ^643,652,653 as illustrated in Figure 1.11: (1) absorption of light, which generates electron–hole pairs in the semiconductor; (2) separation of charge carriers of opposite types; (3) separate withdrawal of these carriers to an external circuit. If one considers a p–n junction under illumination for example, the operation of the semiconducting solar cell can be interpreted as follows: the fundamental feature of the junction is the existence of a strong electric field and in equilibrium the electrochemical potentials on the two sides of the junction are equal, so no net electric current is generated; under illumination, however, electron–hole pairs are generated in the semiconductor and subsequently separated by the electric field effect of the junction. ^654,655

CdTe is an important candidate for solar cell applications because of its high optical absorption coefficient and direct band gap (1.45 eV) that closely matches to the solar spectrum for optimum conversion efficiency. ⁶⁵⁶ For example, a CdTe thin film with a thickness of ∼2 µm can absorb nearly 100% of incident solar radiation. ⁶⁵⁷ Various methods can be utilised for the deposition of CdTe thin films, such as electrodeposition and metal organic chemical vapour deposition. ^658–661 In general, the most common CdTe-based solar cell configuration is composed of a p-type CdTe absorber layer and an n-type CdS window layer. ⁶⁶² A key reason for choosing a CdS/CdTe heterojunction is that the two compounds are miscible and thus an interdiffusion reaction between these two materials will lead to the formation of a CdS_1−x Te _x interlayer, which benefits the solar cell performance. ⁶⁶³ CdS/CdTe thin-film solar cells possess an ideal band gap and a high absorption coefficient meaning that a high energy conversion efficiency above 20% can be achieved. ⁶⁶⁴  

Chalcopyrite Cu(In,Ga)(Se,S)₂—known in shorthand as CIGS—has emerged as one of the most promising photovoltaic materials for low-cost, high-efficiency thin film solar cells. Its utility arises from a compositionally-tuneable band gap, high stability under extended periods of excitation and high optical absorption coefficients resulting from the direct band gap character. ⁶⁶⁵ A high efficiency of 20% for CIGS solar cells has been recorded, ⁶⁶⁶ which highlights them as probably the most viable substitute for Si in solar cells. The band gap in CIGS can approach ideal values by fine tuning the chemical composition; for example, the band gaps of CuInSe₂ and CuGaSe₂ are ∼1.0 and ∼1.7 eV, respectively. ^667,668 Generally, the preparation of CIGS-based solar cells starts with the deposition of an absorber material on a Mo-coated glass substrate. ⁶⁶⁹ The absorber material yielding the highest efficiency is CIGS with a Ga/(Ga+In) ratio of approximately 20% prepared through co-evaporation from elemental sources. ⁴⁴ Overall, the CIGS solar cell has become a relatively mature, highly efficient energy conversion technology that is commercially produced widely. ^670–672  

The hydrogen economy has become a very promising alternative to the current hydrocarbon economy and green hydrogen generation typically involves the process of splitting water into hydrogen and oxygen to harvest renewable energy. The generated hydrogen can then be utilised as a fuel. ⁴²⁸ Metal chalcogenides have arisen as prime candidates for water splitting via catalytic reactions because of highly relevant features such as catalytic edge effects and tuneable band gaps. ^673–676 Such catalytic processes are mainly driven by external electrons from either applied current (electrocatalysis) or light irradiation (photocatalysis).

Typically, in the electrocatalytic process of water splitting, the two half reactions—the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER)—occur at the cathode and anode of an electrolytic cell to produce gaseous molecular hydrogen and oxygen, ⁶⁷⁷ respectively, as illustrated in Figure 1.12. In electrocatalytic water splitting, the HER generally proceeds via two successive steps. The first step of the HER is the adsorption of H via the transfer of a proton and its combination with an electron (the Volmer reaction). The equivalent reaction step in alkaline media requires water dissociation to provide the proton (an energetically more demanding process). For the second step, there are two possibilities: one is the Heyrovsky reaction which tends to occur when the concentration of adsorbed H is relatively low. Here the adsorbed H atom combines with a proton from the electrolyte and an electron transferred from the electrode surface to form one equivalent of H₂; the other possibility is the Tafel reaction, in which two adsorbed H atoms combine with each other to form one molecule of H₂. The determination of which reaction step is rate-determining can be simply discerned by measuring the linear Tafel slope from the polarisation curves. ⁶⁷⁷ The Tafel slope (from a plot of the overpotential vs. log current density) represents the response of the reaction rate (current density) to the potential change. It reflects the difficulty of the electrode process; the higher the slope, the more difficult the electrode process. ⁴²⁸ The overall reaction pathways for the HER are given in eqn (1.7)–(1.10): ^428,678

where the symbol ‘*’ represents the active sites on the catalyst surface, and the ‘ads’ represents adsorbed state of intermediates (H_ads).

Compared to the HER, the anodic OER is a complicated multistep electron–proton transfer process accompanied by the absorption of different oxygen intermediates. ⁶⁷⁹ The generally accepted hypothesis behind the OER mechanism is that the oxygen precursor undergoes the stepwise adsorption–deprotonation–coupling–desorption process (namely, OH_ads → O_ads → OOH_ads → O_2ads), which is typically named the absorbate evolution mechanism (AEM). ⁶⁷⁹ Generally, the AEM is proposed to involve several concerted proton–electron transfer reactions. ⁶⁷⁹ Taking the alkaline OER as an example, OH⁻ is initially chemisorbed on the coordinatively unsaturated metal site and then effectively deprotonates to generate O_ads species (and water + an electron). ⁶⁷⁹ Afterwards, O–O coupling occurs through nucleophilic attack of OH⁻ to form an OOH_ads intermediate. ⁶⁷⁹ Finally, the O₂ molecule desorbs from the catalyst surface along with a second deprotonation (forming water), releasing the active metal site and looping the reaction. ⁶⁷⁹ The overall reaction pathways for the OER in acidic and alkaline media are given in eqn (1.11)–(1.20): ^677,679

where the symbol ‘*’ represents the active sites on the catalyst surface, and the ‘ads’ represents adsorbed state of intermediates (OH_ads, O_ads, OOH_ads, and O_2ads).

The different mechanistic steps of the HER depend on the nature of the electrode materials and the electrolyte media under investigation. ^680,681 Notably, since the dissociation of water is required for HER in neutral and alkaline electrolytes, the reaction in such electrolytes is normally two to three orders of magnitude slower than the equivalent process in an acidic medium. ^682,683 Overall, H₂ production is strongly dependent on the adsorption/desorption processes, which in turn is governed by the corresponding Gibbs free energy. ⁶⁷⁸  

Currently, Ru/Ir-based materials work most efficiently for catalysing the OER and Pt-group metals possess the highest electrocatalytic HER activity. However, the high costs and scarcity of both types of catalyst hinder large-scale applications. ^684,685 Over the past few decades, researchers have devoted intensive efforts to exploring and designing novel OER/HER catalyst candidates with high Earth abundance, reduced manufacturing cost, and excellent performance. ^686–688 Among these candidates, metal chalcogenides are considered as among the most promising electrocatalysts, owing to their low cost and impressive catalytic performance. ⁶⁸⁹ There are generally two strategies to enhance the activity of an electrocatalyst: by improving the intrinsic catalytic activity and by increasing the number of exposed active sites. ⁶⁹⁰ As to the former, approaches such as phase engineering, ^691,692 defect engineering, ^693–695 crystal facet engineering, ⁶⁹⁶ and multi-metal engineering ^697,698 have been developed. Meanwhile, processing a given catalyst by nanostructuring, hybridising with conductive carbon-based materials, and growing on three-dimensional substrates can effectively expose more active sites. ^699–704 These strategies can be combined synergistically to achieve even better catalytic performance.

Among the recently developed OER catalysts based on metal chalcogenides, Ni-based and Co-based chalcogenides have been investigated extensively and exhibit excellent activity and stability. ^705–708 Zhou et al. reported that Ni₃S₂ nanorod arrays had a notable OER activity with a low η ₁₀ value of 187 mV (η ₁₀ is used to evaluate the potential difference between the potential reaching a current density of 10 mA cm⁻² and the thermodynamic potential of HER/OER). ⁷⁰⁷ Gao et al. reported that orthorhombic CoTe₂ had better OER activity and stability compared with hexagonal CoTe, which is ascribed to its large active surface, high electrical conductivity and the low adsorption energy of the *OOH intermediate. ⁷⁰⁸ Furthermore, the incorporation of suitable foreign ions into metal chalcogenides can function well in terms of improving the OER activity. ^431,705 For example, Gu et al. synthesised Fe-doped NiSe₂ ultrathin nanowires with a small η ₁₀ of 268 mV in 0.1 mol L⁻¹ KOH, which was 99 mV and 121 mV lower than that of undoped NiSe₂ and commercial RuO₂, respectively. ⁴³¹ Additionally, the surface properties of metal chalcogenides can be tuned by forming composites with other materials, such that the OER catalytic activity can be enhanced. ^709–712 A representative case is the employment of lamellar CoSe₂ nanobelts, ⁷⁰⁹ which have been investigated in some depth as a platform for the grafting of foreign materials successfully giving rise to high performing OER catalysts. ^710–712 It was observed that chemically grafting an appropriate material (e.g., graphene, Mn₃O₄, and CeO₂) onto CoSe₂ significantly improved its OER activity. ^710–712 For example, the obtained CeO₂/CoSe₂ composite catalyst had a much improved OER activity, with a small η ₁₀ of 288 mV and a small Tafel slope of 44 mV dec⁻¹ in 0.1 mol L⁻¹ KOH, when compared to pure CoSe₂ and RuO₂ catalysts. It was hypothesised that the CeO₂ nanoparticles contain highly mobile oxygen vacancies that could bind adsorbates much stronger than normal oxide sites and additionally act so as to transfer electrons to CoSe₂. Much of the activity of the catalyst is envisaged to originate from the Co(iv) in CoSe₂, which facilitates the formation of OOH and subsequent oxidation to O₂. Therefore, the introduction of CeO₂ appears beneficial in forming OOH_ads on the composite surface and given the high mobility of oxygen vacancies at the nanoscale, promotes water oxidation to produce O₂ on the CeO₂/CoSe₂ composite. ⁷¹²  

The research on chalcogenide HER catalysts begins with MoS₂, which has been identified as an efficient HER catalyst both theoretically and experimentally. ^713,714 The HER activity of MoS₂ arises from exposed edge surfaces that possess an optimised H* adsorption behaviour similar to Pt. ^713–715 This motivated the development of numerous MoS₂ nanostructures to maximise the exposure fraction of edge sites, ^{157,685,692,715–717} leading to a substantially reduced η ₁₀ value of ∼100 mV in acidic electrolytes. However, the HER performance of MoS₂ in alkaline electrolytes is relatively poor due to unfavourable water adsorption and a slow water dissociation step. ^718,719 Doping suitable transition metals (e.g., Ni, Co) into MoS₂ has been shown to accelerate the water dissociation process, thus enabling improved HER activity in alkali electrolytes. ^718–720 Nevertheless, substantial efforts have also been devoted to exploring other metal chalcogenide catalysts for the HER in alkaline electrolytes. ^{429,430,704,721,722} For example, Miao et al. reported a mesoporous FeS₂ material with a η ₁₀ of merely 96 mV and a low Tafel slope of 78 mV dec⁻¹ in 0.1 mol L⁻¹ KOH. ⁷²¹ Alternatively, Yin et al. studied the influence of various metal dopants (Co, Fe and Cu) on the HER activity of NiS₂, and found that Co-doped NiS₂ had the best HER activity and excellent long-term stability. ⁷²² Intriguingly, it has also been reported that P doping can have beneficial effects on dichalcogenide HER catalysts, for example, inducing a phase change of CoSe₂ from cubic to orthorhombic, which resulted in the orthorhombic-CoSe₂|P (8 wt% dopant) exhibiting a much enhanced HER activity (η ₁₀ of 104 mV in 1 mol L⁻¹ KOH) as compared to the cubic form of CoSe₂ (η ₁₀ of 330 mV). ⁴²⁹ Similarly, N doping can have a positive effect, as demonstrated in NiCo₂S₄. ⁴³⁰ The presence of N was proposed to accelerate the desorption of H* from the S sites and promoted water dissociation, delivering a η ₁₀ value of 41 mV—a significant decrease from the value of 104 mV for undoped NiCo₂S₄. ⁴³⁰  

As seen from previous sections, the effective utilisation of solar energy is essential for the achievement of sustainable energy systems. Solar energy collection, conversion and storage are three key processes for practical applications, and these processes can be integrated in a single monolithic photocatalytic water splitting cell, within which solar energy is converted and then stored in simple H–H chemical bonds. ⁷²³ Since Fujishima and Honda's ground-breaking work in 1972 on hydrogen generation utilising an n-type TiO₂ electrode, ⁷²⁴ the research in the photocatalysis field has concentrated on the exploration of suitable catalytic materials. ^725–743 A number of oxide photocatalysts reported show high activities for water splitting, including TiO₂, ^725–730 ZnO, ^731–736 Cu₂O, ^737,738 and WO₃. ^739–743 However, without doping or other optimisation, these oxide photocatalysts are generally active only under ultraviolet light, which accounts for only 5% of the solar spectrum at the Earth surface. ⁷⁴⁴ As a large fraction of solar light is visible light, it is indispensable to explore and develop novel visible-light-driven photocatalysts for water splitting. In this respect, metal chalcogenides have received considerable attention as candidates for visible-light-driven photocatalysts for water splitting, owing to their suitably narrow band gaps and the position of their valence bands at relatively negative potentials compared to oxides. ⁷⁴⁴  

While the source of electrons is supplied by an applied external current in electrocatalytic reactions, photocatalysis is driven purely by light and involves five successive processes, including light-material interaction, photogenerated electron–hole excitation and separation, molecular adsorption, surface redox reaction, and product desorption. ^745,746 When a semiconductor is irradiated by light (where the incident energy is larger than the band gap), electrons and holes are generated in the conduction band (CB) and valence band (VB), respectively. The photogenerated electrons and holes cause redox reactions similar to electrolysis and water molecules are oxidised by the holes to form oxygen and are reduced by the electrons to form hydrogen for overall water splitting. ^744,747 Efficient photocatalytic water splitting relies on the band gap and position of the CB and VB in the electronic structure of the photocatalyst. ⁷⁴⁸ The band gap determines the energy of the absorbed photons and the produced excitons; viz. a narrower band gap implies a wider optical absorption region. In addition, the bottom of the CB must be more negative than the redox potential of H⁺/H₂, while the top of VB should be more positive than the redox potential of O₂/H₂O. ⁷⁴⁹ These criteria indicate that the band gap should be wider than 1.23 eV. ⁷⁴⁹  

Cadmium chalcogenides (such as CdS) are among the most popular visible-light-driven photocatalysts. ^{744,750–753} CdS is a well-established photocatalyst with a band gap of 2.4 eV and demonstrates high activity for hydrogen production under visible-light irradiation. ^751–753 Nanostructured, high surface area cadmium chalcogenides have proved to be extremely effective photocatalysts for several decades. For example, CdS nanoparticles within a narrow particle size envelope (6–12 nm) and with higher hydrogen evolution rates than bulk materials can be prepared by a precipitation process using different zeolite matrices as templates; ⁷⁵² CdS nanostructures containing nanopores with a diameter of ∼3 nm are able to yield a very large BET surface area of ∼112.8 m² g⁻¹ and delivered a hydrogen yield of about 4.1 mmol h⁻¹ under visible-light irradiation. ⁷⁵³ Nanocomposite photocatalysts formed with other metal chalcogenides have also been shown to be successful. ^751,754,755 In one case, MoS₂–CdS and WS₂–CdS nanocomposites could be fabricated from few layers of MoS₂ or WS₂ grown upon the surface of CdS, and the resulting nanohybrids rapidly produced astonishingly high yields of hydrogen (1472 mmol h⁻¹ g⁻¹ for MoS₂–CdS, 1984 mmol h⁻¹ g⁻¹ for WS₂–CdS) suggesting that the surface of the nanocomposites contained a higher concentration of active sites than CdS itself. ⁷⁵⁵ In fact, MoS₂ and WS₂ are increasingly finding application more widely as components in composites and heterostructures formed with photocatalytically-active oxides such as TiO₂ and other chalcogenides such as ZnIn₂S₄. By forming the composites increased visible light absorption and enhanced charge carrier mobility have been observed. ^756,757 All these efforts indicate the immense potential of non-precious metal chalcogenide-based catalysts in terms of the photocatalytic production of hydrogen.

This chapter aims to serve as a valuable introduction to the diverse and exciting contemporary solid-state materials chemistry of chalcogenides. In the opening sections it seeks to summarise the basic structures and properties of metal chalcogenides, focuses on the structural characteristics of typical layered metal chalcogenides, briefly captures the development of nanostructured metal chalcogenides and then introduces MoS₂ and Bi₂Te₃, two typical chalcogenide materials with a wide range of applications. The chapter subsequently considers some of the principal synthesis methods that are available to produce single crystals, polycrystalline bulk powders, thin films and nanostructures of metal chalcogenides. Finally, the chapter concentrates on some of the most important fields of development in chalcogenide chemistry today, emphasising the energy storage and energy conversion applications of metal chalcogenides, including the basic principles behind these technologies, some of the foremost metal chalcogenides in use and some of the properties that lie behind their performance.

Chalcogenides with diverse structures and rich compositions exhibit an intriguing mix of electronic, physical and chemical characteristics which not only lead to unusual and rare phenomena but also carve a path to useful applications. There is still plenty of scope for the further manipulation of such a plentiful array of chemical and physical properties and there are many chalcogenide systems that are yet to see their remarkable features exploited. Behind such developments, the understanding of the function of chalcogenides and the underpinning chemistry responsible is critical. We hope that this introduction to some of the fundamentals of chalcogenide chemistry—including crystal chemistry, synthesis approaches and key properties and applications—might inspire the reader to read, learn further and to go on to discover still more regarding these fascinating materials.