1 Introduction
Recent years have witnessed the great breakthrough of semiconductor technology due to the introduction of silicon transistors [
1]. Following the Moore’s law, the number of transistors in chips is exponentially increasing and the device performances are improved significantly, in consideration of the continuous miniaturization of silicon-based process nodes [
2,
3]. However, the Moore’s law is becoming unsustainable due to the physical limits of bulk silicon materials, the short-channel effects and high heat dissipation phenomena are inevitably appeared, and further reducing the size of silicon-based field-effect transistors (FETs) is facing a major challenge [
4]. In this context, two-dimensional (2D) materials, such as graphene [
5,
6], boron nitride [
7,
8], and transition metal dichalcogenides (TMDCs) [
9,
10], have attracted extensive attention due to their novel physical properties and excellent device performances, such as anomalous Hall effect [
11], room temperature quantum Hall effect [
12], ultrahigh carrier mobility [
13], excellent thermal conductivity [
14], and light transmission [
15]. The sizes of FETs can be further reduced by replacing the silicon channels with 2D materials [
16,
17].
As a unique member of 2D materials, the semiconducting TMDCs have exhibited great potentials in the state-of-the-art FETs [
18]. The layered structure allows the channel to be thinned to the monolayer and exhibits excellent electrostatic control capacity, and thus the operating voltage/current and energy consumption can be reduced effectively [
19,
20]. The smooth surface without dangling bond reduces the degradation of carrier mobility induced by the scattering [
21,
22]. The abundant band structure enables the design of novel logic and memory device [
23,
24]. For example, by using the exfoliated few-layer and monolayer MoS
2 nanosheets as the channels, the ultrahigh electron mobility (~34 000 cm
2·V
−1·s
−1) [
25] and excellent on/off current ratio (~10
8) [
26] was obtained, respectively. The ultralow contact resistance (~0.2 kΩ·μm) was achieved in monolayer MoS
2 transistor by constructing the 1T-2H MoS
2 lateral heterostructure [
27]. By using a physically assembled silicon nanowire as the lift-off mask, the ultrashort channel device with few-layer MoS
2 channel was created and showed remarkable conduction current density (~0.83 mA·μm
−1) at room temperature [
28]. The abovementioned results reveal that the 2D semiconducting TMDCs offer a promising playground for the construction of high performance electronic devices.
However, the device performances of 2D semiconducting TMDCs are still unsatisfactory, and which are limited by the following bottlenecks. First, the intrinsic defects (e.g., vacancies, grain boundaries, and impurities, etc.) in 2D TMDCs increase the carrier scattering sites and reduce the device performances [
29-
32]. Second, although the band structures of 2D TMDCs can be modulated and new physical phenomena can be introduced by means of a doping strategy [
33-
37], the excessive doping should bring in the deep impurity levels and reduce the device performances [
38,
39]. Third, the difference between the work function of metal electrodes and the electronic affinity energy of 2D semiconducting TMDCs results in high Schottky barrier height and contact resistance [
40-
43]. Therefore, the controlled synthesis of high-quality samples with few defects, the development of ingenious doping strategy, and the optimization design of interfacial contacts between electrodes and channels are crucial for improving the device performances of 2D semiconducting TMDCs.
This article hereby aims to provide an up-to-date summary or discussion on the aforementioned issues as described in Fig.1. Specifically, the chemical vapor deposition (CVD) synthesis of wafer-scale 2D semiconducting TMDCs single crystals will be introduced, and the internal mechanisms will be discussed. The accurate doping of 2D semiconducting TMDCs by the metal atoms and organic molecules will be reviewed. Furthermore, some new methods including the electrode transfer, the introduction of semimetal electrodes, and the construction of 2D metallic/semiconducting TMDCs heterostructures will also be summarized to optimize the interfacial contact.
2 Controllable synthesis of wafer-scale 2D semiconducting TMDCs single crystals
The ultrahigh carrier mobility, excellent on/off current ratio, and distinguished subthreshold swing of 2D semiconducting TMDCs provide a wide application prospect in the high performance electronic devices [
5,
44-
46]. However, the presence of grain boundaries in polycrystalline TMDCs films increases the carrier scattering sites and reduces the carrier mobilities and conductance [
47-
49]. Therefore, controllable synthesis of wafer-scale 2D semiconducting TMDCs single crystals is essential for the electronic device applications. Recently, fruitful achievements have been attained in growing 2D single crystals [
50–
56]. For example, inch-sized graphene single crystals were synthesized on Cu−Ni alloys by controlling the nucleation [
50]. Through the seamless stitching of unidirectional domains, wafer-scale 2D single crystals were obtained, such as graphene on Ge(110) [
51], h-BN on liquid Au substrate [
54], vicinal Cu(110) [
55], and Cu(111)/sapphire [
56]. Notably, there are two distinct routes for synthesizing wafer-scale 2D single crystals: (i) the nucleation controlling: allowing only one nucleation site on a wafer-scale substrate, and then the single nucleus further grows to a wafer-scale single crystal; (ii) the orientation controlling: a substantial number of unidirectional domains epitaxially grow on a single-crystalline substrate, and further generate a wafer-scale single crystal by the seamlessly stitching [
57].
2.1 Synthesis of 2D semiconducting TMDCs single crystals by the nucleation controlling
The nucleation controlling is extremely difficult for synthesizing wafer-scale 2D semiconducting TMDCs single crystals in view of the complicated growth kinetics. Nevertheless, the phase transition design is a plausible method to solve such an issue. Xu
et al. [
58] developed a solid-to-solid phase transition and recrystallization strategy to grow 2D semiconducting 2H-MoTe
2 single crystals on amorphous insulating substrates. As illustrated in Fig.2(a), a single-crystalline 2H-MoTe
2 nanoflake was deliberately implanted on the center of 1T′-MoTe
2 wafer and used as the seed crystal to trigger the phase transition and recrystallization. Subsequently, a dense Al
2O
3 film (~30 nm) was deposited on the wafer to isolate the 1T′-MoTe
2 from Te precursors and prevent the random nucleation of 2H-MoTe
2. A small hole was introduced into the seed region and served as the only channel of Te precursors supply.
Notably, during the annealing process, the Te vacancies were reduced and the rearrangement of Mo and Te atoms of the bottom 1T′-MoTe
2 layers triggered the transition from 1T′ to 2H phase. The lattice structures, crystal orientations, and stacking orders of the bottom MoTe
2 were identical to the seed crystal after the recrystallization process, as proved by the aberration-corrected scanning transmission electron microscopy (AC-STEM) in Fig.2(b). Interestingly, with increasing the annealing time, the phase transition expanded to the outside seed regions and then the wafer-scale 2H-MoTe
2 single crystal was formed, as shown in Fig.2(c) and (d). In addition, a large-scale coplanar heterophase of 1T′ /2H/1T′-MoTe
2 FET arrays were then fabricated to verify the device performance uniformity of such a single crystal film [Fig.2(e)]. As shown in Fig.2(f) and (g), the excellent room temperature carrier mobility (45 ± 2 cm
2·V
–1·s
–1) and on/off current ratio [(1.8 ± 0.3) × 10
4] were obtained, comparable to the reported values of single-crystalline 2H-MoTe
2 nanoflake transistors. And such results suggested the ultrahigh device performance and its spatial uniformity of 2H-MoTe
2 single crystal [
59,
60].
In short, the phase transition design provides an innovative route for synthesizing wafer-scale 2D semiconducting 2H-MoTe2 single crystals. However, the uniform and atomically flat few-layer 1T′-MoTe2 film is required to serve as the precursor of phase transition, and it is difficult to synthesize monolayer 2H-MoTe2 single crystal. Moreover, this method is also inapplicable for the growth of other 2D TMDCs with large energy differences between different phases.
2.2 Synthesis of 2D semiconducting TMDCs single crystals by the orientation controlling
The orientation controlling was proven to be a reasonable method for the direct synthesis of wafer-scale 2D single crystals [
51,
54]. However, it remains challenging to obtain the unidirectional domains of 2D TMDCs due to the symmetry mismatch between TMDCs and substrate [
57]. Chen
et al. [
61] proposed that the parallel steps of sapphire (0001) surfaces could be served as the nucleation sites of monolayer WS
2 and further induced the formation of unidirectional alignment domains. Recently, Li
et al. [
62] realized the epitaxial growth of 2-inch monolayer MoS
2 single crystal on a
c-plane sapphire with a miscut angle towards
a-axis. As presented in Fig.3(a) and(b), the parallel steps of
c/
a sapphire were used as the nucleation sites for growing monolayer MoS
2 with unidirectional alignments and single crystal films [Fig.3(c) and (d)].
The first-principles calculations were then performed to clarify the growth mechanism, as revealed in Fig.3(e). The presence of parallel steps lowered the symmetry of
c-plane sapphire and broke the degeneracy of formation energy of two possible edge configurations (ZZ−Mo−S
2 and ZZ−S
2), which resulted in the most energetically favorable configuration of ZZ−Mo−S
2 for the unidirectional growth of monolayer MoS
2. Furthermore, the monolayer MoS
2 FET arrays were also fabricated to evaluate the device performance and its uniformity, as presented in Fig.3(f). The statistical results of 151 monolayer FETs showed that both the carrier mobilities and subthreshold swings subjected to the Gaussian distribution with the variation of ~15% and ~11%, respectively, suggested the high device performance uniformity of wafer-scale monolayer MoS
2 single crystal [Fig.3(g)]. The average and maximum carrier mobility was obtained to be ~77.6 and ~102.6 cm
2·V
–1·s
–1, respectively, much higher than that of the monolayer polycrystalline MoS
2 film [
63,
64]. Notably, the terrace height of sapphire substrate is also crucial for synthesizing wafer-scale bilayer MoS
2 [
65,
66].
In addition, as shown in Fig.3(h), the orientation controlling method was also employed to synthesize wafer-scale monolayer WS
2 single crystals on vicinal
a-plane sapphire [
67]. The low-energy electron diffraction (LEED) pattern captured from different regions of WS
2 films showed two sets of 3-fold symmetry spots with different diffraction intensities [Fig.3(i)], confirmed the formation of monolayer WS
2 single crystal. Furthermore, the parallelly aligned holes with hexagonal shapes were generated on monolayer WS
2 film during the oxygen etching process, reconfirmed the epitaxial growth behavior and seamless stitching of misorientation-free WS
2 domains on the vicinal
a-plane sapphire [Fig.3(j)]. Besides the step-guided strategy, the other methods were also proposed to regulate the orientations of 2D semiconducting TMDCs and synthesize the wafer-scale single crystals. For example, the parallelly aligned monolayer MoS
2 domains were formed as increased the S/Mo precursor ratio, in view of the complete sulfurization of Mo precursors [
68,
69]. In addition, the growth temperature is also crucial for growing wafer-scale 2D semiconducting TMDCs single crystals, and the high temperature results in the undulation of step alignment and the misorientation of TMDCs domains [
70].
In short, by the nucleation and orientation controlling, the wafer-scale 2D semiconducting TMDCs (e.g., MoS2, WS2, 2H-MoTe2) single crystals have been successfully synthesized and which lays the solid foundation for constructing high performance electronic devices. Nevertheless, the controlled growth of 2D semiconducting single crystals remains challenging, such as the complicated growth process of phase transition and recrystallization, as well as the expensive growth substrates of c/a and a-plane sapphire.
3 The accurate doping of 2D semiconducting TMDCs
The controlled doping of 2D semiconductors is an important strategy for modulating their electronic properties and introducing novel physical phenomena [
71–
74]. For example, the band structure, carrier density, and photoluminescence intensity of 2D semiconducting TMDCs can be tuned by means of accurate doping. However, the excessive doping should introduce the deep impurity levels and reduce the device performances [
38,
39]. In this section, we will summarize the recent experimental advances regarding the accurate doping of 2D semiconducting TMDCs and its influence on the device performances.
3.1 The Fe-doping of monolayer MoS2
As a type of interesting dopants, the magnetic transition metal atoms (e.g., Fe, Co, and Mn, etc.) could introduce the novel physical phenomena in 2D TMDCs [
75–
78]. Recently, Li
et al. [
79] reported the CVD synthesis of large-area monolayer Fe-doped MoS
2 on SiO
2/Si. As shown in Fig.4(a), the monolayer Fe-doped MoS
2 triangles were obtained with the maximum domain size of ~250 µm. Notably, the Fermi level of monolayer MoS
2 was moved toward the conduction band minimum (CBM) after the Fe doping due to the additional bands from
s and
p orbitals of Fe, indicated the n-type doping behavior, as revealed in Fig.4(b).
To evaluate the device performance, the monolayer Fe-doped MoS2 FETs were thus constructed by using Cr/Au as the electrodes. Excellent room temperature on/off current ratio and electron mobility (~108 and ~54 cm2·V–1·s–1) were obtained for monolayer Fe-doped MoS2, much higher than those of the pristine monolayer MoS2 (~107 and ~6 cm2·V–1·s–1), suggested the enhanced device performance. Furthermore, the contact resistance was also extracted by using the transfer length method (TLM), as shown in Fig.4(c). The contact resistance of monolayer Fe-doped MoS2 was calculated as ~678 Ω·µm, much smaller than that of the pristine monolayer MoS2 (~117 kΩ·µm), indicated the outstanding interfacial contacts between monolayer Fe-doped MoS2 and electrodes. To clarify the internal mechanism, the temperature-dependent electrical measurements were thus performed. The energy barriers of monolayer Fe-doped MoS2 and pristine monolayer MoS2 were extracted from the Arrhenius plots in Fig.4(d), and the perfect Ohmic contact was confirmed for Fe-doped MoS2. Furthermore, the enhanced electron mobility was also observed in monolayer Fe-doped MoS2 as the temperature decreasing from ~300 K to ~100 K, indicated the suppressed ionized impurity scattering [Fig.4(e)]. Such results reveal that the device performances of 2D semiconducting TMDCs can be improved by the accurate Fe doping, and which provides a new scheme for downscaling the electronic devices.
3.2 The remote modulation doping of few layer MoS2 by triphenylphosphine
Although the doping is required for modulating the electronic properties of semiconductors, the introduction of impurities should result in the Coulomb scattering and hinders the charge transport, especially for 2D semiconductors. To avoid this issue, a remote modulation doping strategy was developed in the WSe
2/h-BN/MoS
2 heterostructures [
80]. The triphenylphosphine (PPh
3) was selected as the dopants and then imposed on the top layers of WSe
2 [Fig.5(a)]. Notably, the electrons from PPh
3 were introduced into WSe
2 layers through the molecular doping and spontaneously transferred to MoS
2 layers in view of the conduction band offset, which resulted in the n-type doping in MoS
2 channel, as shown in Fig.5(b).
To verify the doping capability of PPh3, the electrical measurements of modulation-doped (MD) WSe2/h-BN/MoS2 and directly doped (DD) MoS2 FETs were performed. The negative shifts of threshold voltages were clearly observed for both MoS2 devices after the PPh3 treatments, indicated the n-type doping behavior, as shown in Fig.5(c). The same variation tendency was also demonstrated in the electron density, which suggested that the PPh3 molecules could modulate the carrier density of underlying MoS2 channels through the remote charge transfer across the WSe2/h-BN layers without a substantial loss in the doping efficiency. Notably, the electron densities of such two type MoS2 devices could be tuned by the doping concentrations and treatment time. In addition, the carrier mobility of MoS2 in MD device remained unchanged at ~60 cm2·V–1·s–1, even after the electron doping. Nevertheless, a degraded mobility from ~60 cm2·V–1·s–1 to ~35 cm2·V–1·s–1 was obviously observed for the DD counterparts [Fig.5(d)]. Such results suggested that the charged impurity scattering was suppressed through the remote modulation doping strategy.
3.3 The photo-induced and electrostatic doping strategies of 2D TMDCs
In addition, the other doping methods, such as the photo-induced and electrostatic doping, have also been proposed in 2D TMDCs. For example, Seo
et al. [
72] reported a reversible photo-induced doping of few-layer 2H-MoTe
2 and 2H-WSe
2 as revealed in Fig.6(a)–(c), and the polarity of channel was reconfigured from n-type to p-type under the different laser light frequencies. The light-lattice interactions, as well as the self-interstitial defects of Te under ultraviolet illumination and the subsequent substitutions of O in Te/Mo vacancies under visible illumination resulted in the reconfigurable doping effect. By using such a method, the complementary metal-oxide-semiconductor (CMOS) device on a single channel was also constructed, and the circuit functions were dynamically reset from the inverter to the switch under the different light frequencies, as shown in Fig.6(d). Furthermore, Wang
et al. [
81] reported an electrostatic-doping-driven phase transition between the hexagonal and monoclinic phases of monolayer MoTe
2, as demonstrated in Fig.6(e) and proved by the hysteretic loop in Raman spectra [Fig.6(f)]. The electrostatic doping method opens up a new possibility for the construction of phase-change devices at the atomic scale.
Briefly, the band structures and electronic properties of 2D semiconducting TMDCs can be modulated by means of ingenious doping, and which provides a new paradigm for improving the device performances and extending the Moore’s law. However, the internal mechanism of metal doping should be further explored and the stability of remote modulation doping should also be enhanced.
4 The optimization design of interfacial contacts between electrodes and 2D semiconducting channels
The optimization design of interfacial contacts between electrodes and channels is crucial for realizing the applications of 2D semiconductors in high performance electronic and optoelectronic devices [
41,
43,
82–
84]. Notably, the interfacial contact quality can be evaluated by the contact resistance. For example, the low contact resistance is crucial for achieving high “on” current density, large photoresponse, and high-frequency operation [
85]. Nevertheless, the difference between the work function of metal electrodes and the electronic affinity energy of 2D semiconducting TMDCs results in the formation of Schottky barrier [
86]. Furthermore, the Fermi-level pinning effect is also induced by the disorder-induced gap state (DIGS) and metal-induced gap state (MIGS) [
87], which breaks the Schottky−Mott rule. Therefore, the optimization design of interfacial contacts is significant for improving the device performances of 2D semiconducting TMDCs. In this section, we will review three methods to optimize the interfacial contacts including the electrode transfer, the introduction of semimetal electrodes, and the construction of 2D metallic/semiconducting TMDCs heterostructures.
4.1 The electrode transfer for improving the device performances of 2D semiconducting TMDCs
The defects and disorders should be introduced on the surfaces of 2D semiconducting TMDCs during the lithography and thermal evaporation processes, which degenerate the device performances accordingly. Notably, the van der Waals contacts between electrodes and 2D semiconductors without Fermi-level pinning are theoretically possible. For example, the clean interfaces were created in the van der Waals metal−semiconductor junctions by laminating the metal electrodes with atomically flat surfaces onto the dangling-bond-free 2D semiconductors, as shown inFig.7(a) [
88]. The excellent device performances (with the electron and hole mobility up to ~260 and ~175 cm
2·V
–1·s
–1, respectively) were clearly observed in the exfoliated few-layer MoS
2 by using the transferred Ag and Pt electrodes [Fig.7(b), (c)]. Furthermore, the
S factor, which was characterized the Fermi level pin strength, was obtained to be ~0.96 for the devices with transferred metal electrodes, approached to the Schottky−Mott law limit defined by the electrostatic energy alignment. On the contrary, the
S factor was calculated as ~0.09 for the control devices with evaporated metal electrodes, indicated strong Fermi-level pinning effect near the CBM at the electrode/MoS
2 interfaces, as demonstrated in Fig.7(d). Such an interesting electrode transfer establishes a highly efficient and damage-free strategy for constructing high performance electronic devices.
In addition, the interaction- and defect-free van der Waals contacts were formed between the metal electrodes and exfoliated few-layer WSe
2 nanosheets via a metal deposition process that used the Se buffer layer [
89]. In detail, the Se layer was used to protect the WSe
2 channel during the metal evaporation process and it could be removed by the annealing process at ~150 ℃ [Fig.7(e)]. The n-type behavior was clearly observed for the direct Au contacted few-layer WSe
2 device, in view of the Fermi-level pinning effect near the conduction band at the Au−WSe
2 interface. However, the p-type characteristic with a large on/off current ratio (>10
6) and high hole mobility (~135 cm
2·V
–1·s
–1) was presented in the van der Waals contacted few-layer WSe
2 due to the perfect interface [Fig.7(f)]. Density functional theory (DFT) calculations were then performed to understand the band structure at the Au−WSe
2 interface [Fig.7(g), (h)]. The new states were formed in the mid-bandgap region of the direct Au contact WSe
2, and changed the band bending direction. Nevertheless, the density of states (DOS) of WSe
2 were formed in the van der Waals contact with Au electrodes and exhibited few mid-gap states. The construction of van der Waals contact using a Se buffer layer provides new route for controlling the Schottky barrier height and can be used for fabricating the large-scale electronic devices.
4.2 The introduction of semimetal electrodes for improving the device performances of 2D semiconducting TMDCs
The Fermi-level pinning effect can be suppressed by constructing van der Waals contact, nevertheless, the tunneling barriers are inevitable because of the interface gaps between electrodes and channels. In addition, the electrode transfer process should be implemented in a sufficiently clean environment, therefore, this approach is vulnerable to the external influence and difficult to apply to industrially relevant large-substrate technology. Herein, the exploitation of new electrode materials is necessary.
Notably, as the semiconductor closed to a metal surface, the extended wavefunction of metal perturbed the environment of semiconductor and resulted in the emergence of MIGS [Fig.8(a)], totally different from the initial state of semiconductor [Fig.8(c)] [
90]. Specially, if the Fermi level of electrode lied inside the semiconductor bandgap, the Schottky barrier was thus formed unavoidably. However, for the semimetal Bi electrode, the zero DOS near the Fermi level induced few MIGS. Particularly, its Fermi level closed to the CBM of monolayer MoS
2, and the conduction-band-contributed MIGS were greatly suppressed [Fig.8(b)]. As a result, the degenerate state and free of Schottky barrier at the interface were formed as the semimetal Bi contacted monolayer MoS
2. The schematic diagram of monolayer MoS
2 back-gated FET with Bi contact was shown in Fig.8(d). The ultralow contact resistance (~123 Ω·μm at the carrier density of 1.5×10
13 cm
−2) and zero Schottky barrier height was obtained, respectively, as revealed in Fig.8(e) and (f), indicated the ultrahigh device performance of Bi contacted monolayer MoS
2. Notably, the contact resistance values were comparable to those of three-dimensional semiconductors, and approached to the quantum limit [Fig.8(g)]. Even so, in view of the low work function of Bi electrode, it is incapable to reduce the Schottky barrier height of hole. Therefore, searching for high work function semimetal electrodes is crucial for fabricating high performance p-type devices.
4.3 The construction of 2D metallic/semiconducting TMDCs heterostructures
2D metallic TMDCs have attracted great attention due to their interesting physical properties (e.g., charge density waves order, superconductivity, and magnetism, etc.) [
91–
108]. In addition, some revolutionary applications such as transparent electrodes and energy conversion/storage have also explored in the 2D metallic TMDCs, by virtue of the ultrahigh conductivities, excellent electrocatalytic activities, and novel layered structures [
109–
124]. Notably, the 2D metallic TMDCs are considered as ideal electrode materials for improving the device performances of 2D semiconducting TMDCs in view of their similar atomic structures and complementary electronic properties [
27,
59,
125–
133].
The VS
2−MoS
2 lateral heterostructures were successfully synthesized on SiO
2/Si by a two-step CVD method, as revealed in Fig.9(a) and (b) [
134]. The back-gate monolayer MoS
2 FETs were then fabricated with the lateral VS
2 and vertical Ni contact, respectively [Fig.9(c)]. The contact resistances of such two typical monolayer MoS
2 devices were extracted by the four-probe measurements and plotted in Fig.9(d). Notably, the contact resistance of VS
2−MoS
2 was obtained to be ~520 Ω·μm, more than ten times smaller than that of Ni-MoS
2 contacts (~8640 Ω·μm) [Fig.9(d)], indicated the perfect interfacial contact. Furthermore, the 2D metallic/semiconducting TMDCs heterostructure arrays were also CVD synthesized by using a focused laser irradiation and raster scan method, as shown in Fig.9(e) [
135]. The VSe
2/WSe
2 heterostructure array FETs were thus constructed [Fig.9(f)], and demonstrated ultrahigh device performances with the mobility of ~135 cm
2·V
–1·s
–1, much higher than that of Cr/Au contacted monolayer WSe
2 (~10 cm
2·V
–1·s
–1), which was attributed to the protective effect of VSe
2 on the monolayer WSe
2 channel during the electrode evaporation process [Fig.9(g)].
5 Conclusion and perspectives
2D semiconducting TMDCs have shown great application potentials in the electronic devices due to their novel physical properties, atomically thin thicknesses, and abundant band structures. However, the device performances are still lower than that of the theoretical values, which is attributed to the intrinsic defects, excessive doping, and daunting contacts. From these perspectives, we have summarized three strategies for improving the device performances of 2D semiconducting TMDCs, including (i) the synthesis of wafer-scale 2D semiconducting TMDCs single crystals by the nucleation and orientation controlling to reduce the grain boundary densities, (ii) the accurate doping of 2D semiconducting TMDCs by the metal atoms and remote modulations to tune the band structures and suppress the impurity scatterings of 2D semiconducting TMDCs, and (iii) the optimization design of interfacial contacts between electrodes and 2D semiconducting TMDCs channels by means of transferring metal electrodes, introducing semimetal electrodes, and constructing 2D metallic/semiconducting TMDCs heterostructures.
Although the impressive achievements have been made in such aspects, some daunting challenges are still unresolved. For the synthesis of wafer-scale 2D semiconducting TMDCs single crystals, the high-cost sapphire substrates with a small miscut angle are used for controlling the domain orientations. Moreover, the longtime pretreatment of substrates at extremely high temperature is needed before the synthesis of 2D TMDCs. Thus, it is urgent to develop a facile technology for growing wafer-scale 2D semiconducting TMDCs single crystals on the commercially available substrates. For the controllable doping, the organic molecules are unstable in the atmosphere, and which are incompatible with the device construction. Furthermore, the intrinsic mechanism of metal doping is still unclear, and more experimental and theoretical explorations are required. For the optimization design of interfacial contact, the electrode transfer method is complicated, and the gap between electrode and channel will introduce the tunneling barrier. In addition, the small work function of semimetal electrode is disadvantageous to reduce the Schottky barrier height of p-type semiconductor. The controllable synthesis of wafer-scale 2D metallic/semiconducting TMDCs heterostructures is still unaddressed.
Although there are some challenges, many breakthroughs will be made in the near future regarding the improvement of device performances of 2D semiconducting TMDCs. For instance, the interactions between 2D TMDCs domains and substrates can be tuned by the metal doping during the CVD growth process, and it is feasible to obtain the wafer-scale 2D semiconducting TMDCs single crystals on the commercially available substrates. Meanwhile, the metal doping is also beneficial for modulating the band structures of 2D TMDCs and reducing the contact resistances between electrodes and channels. In addition, some newfangled physical properties, such as magnetism, can be introduced by the metal doping, and which provides a new route for the construction of multifunctional electronic devices.