Chemical vapor deposition for few‐layer two‐dimensional materials

Chemical vapor deposition (CVD) approach offers a controllable strategy for preparing large‐area and high‐quality few‐layer (mainly bilayer or trilayer) twisted or untwisted two‐dimensional (2D) materials, and is predicted to boost the development of 2D materials from laboratory research to industrial applications.

Researchers have been captivated by the properties of bilayer/trilayer two-dimensional (2D) materials since the discovery of magic-angle twisted bilayer graphene's (tBLG) superconducting characteristic. 1,2 Bilayer/trilayer 2D materials, both twisted and untwisted, have essential applications in nanoelectronics, nano-optics, catalysis, and anti-oxidation. [3][4][5][6] The investigation on bilayer/ trilayer 2D materials obtained through artificial stacking exfoliated layers, on the other hand, is limited to laboratory research because of the difficulty in scaled production and the facility to introduce interlayer contamination, which severely restrict their further wide applications. Chemical vapor deposition (CVD) offers compelling benefits in mass production and precise controllability. However, breaking through the energy preference in the atomic deposition process to synthesize few-layer 2D materials with desired spatial structures still remains a great challenge.
In a recent Article of Nature Communications, Zhongfan Liu's group proposed a hetero-site nucleation strategy for increasing the proportion of tBLG in all bilayer graphene products from 16% to 88%. 7 The nucleation of the second-layer graphene arises at a separate site owing to a gas-flow perturbation caused by a sudden increase of H 2 and CH 4 as shown in Figure 1A. Both the carbon-isotope-labelling experiments and computational fluid dynamics show that the shift distance between the growth center of the second layer of graphene from that of the first layer is positively correlated with the timing when the perturbation is provided. The nucleation of the second layer is preferred to occur close to the edge of first-layer graphene, because there exists a large energy barrier for carbon atoms diffusing from the Cu surface to graphene-covered region. The two layers then continue to grow around the nuclei and the differing local environment around two nuclei finally results in production of tBLG. In addition, the proportion of bilayer graphene is determined by the ratio of hydrogen before and after the sudden increase, and it achieves its maximum when the hydrogen partial pressure reaches the value of 2.5 times as that of before after a sudden increase. The proportion would also rise as the CH 4 flow rate after the perturbation increases; however, an excess partial pressure of H 2 or CH 4 induces the generation of undesirable multilayer graphene, lowering the proportion.
Selective area electron diffraction (SAED) and highresolution transmission electron microscope (HRTEM) were conducted to probe the twist angle of bilayer graphene. It is worth mentioning that while producing tBLG with a tiny twist angle (<3°) is challenging, the proportion of the tBLG with twist angles approaching 30°i s substantially higher than other tBLG. HRTEM patterns reveal clear Moiré periods, suggesting the successful creation of a high-quality twisted structure and quasicrystalline system. Furthermore, the electrical performance of the Hall bar device based on tBLG was investigated, and ultrahigh room-temperature carrier mobilities of 67,000 and 68,000 cm 2 /(V·s) were found for electrons and holes, respectively. This work inspires the controllable preparation of twisted bilayer 2D materials with large interlayer rotating angles. In another work published in a recent issue of Science, Hyeon Suk Shin and their colleagues reported an epitaxial growth of wafer-scale single-crystal trilayer untwisted hexagonal boron nitride (hBN) on Ni(111) substrate. 8 The formation process of wafer-scale hBN can be described as the coalescence of multiple trilayer hBN islands and the growth mechanism is quite different from previously reported precipitation from the bulk. Their findings demonstrate that the growth of trilayer hBN is mainly mediated by surface. Noting that only a temperature range of 1120-1270°C can produce uniform and continuous hBN, which is substantially higher than the growth temperature of around 1000°C used in past. In addition, both SAED and HRTEM characterization results suggest that Ni 23 B 6 emerges during the cooling process, and the trilayer hBN presents a stacking configuration of AA′A.
To further investigate the growth mechanism of trilayer hBN, they first compared the binding energies of N 7 B 6 (commonly regarded as hBN nucleus) on the terrace or step edge at rotation angles of 0°or 60°, respectively. The simulation results indicate that the 0°-rotation-angle N 7 B 6 at a terrace is the most energetically preferred. Furthermore, an untwisted structure is guaranteed because the van der Waals interlayer interactions only exhibit locally minimum at 0°or 60°( Figure 1B), the occurrence of other twists necessitates crossing huge energy barriers. Thus, all hBN islands emerge at the same direction, even if grow across edge steps, ensuring the formation of wafer-scale single-crystal hBN film on Ni(111) substrate without any twists. They verified the outstanding performance of the single-crystal trilayer hBN at the end of their article. When used to defend against electrochemical degradation, it showed no signs of decline after 2000 cycles, whereas polycrystalline hBN could only endure 50 cycles of peer-off tests. Single-crystal untwisted trilayer hBN, on the other hand, when used as a gate dielectric in a field effect transistor, can clearly blocked charge trapping and prevented the electron doping from the SiO 2 substrate, improving the mobility of semiconductor by up to 60.7%. Their work makes substantial contributions to the preparation of wafer-scale untwisted few-layer 2D materials.
In addition to bilayer graphene and trilayer hBN, fewlayer transition metal dichalcogenides (TMDs) or vertical heterostructures based on TMDs have also received extensive attentions. Shu Ping Lau et al. recently developed a scalable one-step CVD method for untwisted heterobilayers of TMDs and demonstrating ultrathin devices with unanticipated out-of-plane ferroelectric and piezoelectric characteristics. 9 Their CVD-prepared commensurate MoS 2 /WS 2 heterobilayers can be assigned to a unique point group (3 m) of their own, and have no vertical symmetry to neutralize the out-of-plane strain, unlike monolayer or homobilayer TMDs, therefore, leading to out-of-plane features.
According to their findings, MoS 2 /WS 2 heterobilayers energetically prefers to stack with a rotation angle of 0°( 3R-like structure) or 180°(2H-like structure), which results in neither twist angles nor Moiré domains. The out-of-plane piezoelectric constants of 3R-like and 2Hlike heterobilayers were studied using piezoresponse force microscopy maps, and it was discovered that the piezoelectric constant of 3R-like heterobilayers was slightly higher than that of 2H-like heterobilayers ( Figure 1C), and was comparable to the in-plane piezoelectric constant of monolayer MoS 2 . The internal polarization changed by external electric field may result in a switching between two stable states of piezoelectric and ferroelectric. When the electric field is raised from 1.8 to 2.7 V, the MoS 2 /WS 2 heterobilayer exhibit a roomtemperature ferroelectric property since a typical Vshape piezoresponse hysteresis loop can be observed in the heterobilayer-based device. Their work challenges the conventional belief that commensurate structures only exhibit in-plane ferroelectricity and piezoelectricity unless twist angles or Moiré domains are present, and it offers a creative idea for producing out-of-plane piezoelectric and ferroelectric 2D materials simply using CVD method.
Much work remains to be done in developing controllable CVD methods for desired few-layer 2D materials, such as tBLG with specific twist angles, twisted hBN, and TMDs, the heterostructures with controllable layer number and tailored superlattices. Delightfully, very recently, Kaihui Liu et al. have suggested a designed growth of large bilayer graphene with arbitrary twist angles through the usage of prerotated Cu(111) foils. 10 Their work considerably inspires more researchers to explore vigorously. Despite the fact that CVD was previously used as a relatively controllable method to create monolayer 2D materials, it is now necessary to constantly adapt to advanced requirements and enhance the controllability to produce the few-layer 2D materials listed above due to the rising demand for novel physical properties and nextgeneration electronics.