Two-dimensional (2D) materials are a kind of monolayer or fewer layer materials whose thickness is much smaller than the lateral dimension. Many of them have novel electronic^[1-4] and optoelectronic properties^[5-7], such as, the quantum-confined electronic band structures^[8], layer-dependent bandgaps^[9], and atom-lattice-related conductance transition from semiconductors to metals^[10] As a classic 2D material, graphene has excellent electron mobility with excellent performance in transistors and as electrochemical electrodes^[11]. However, one weakness of the electronic structure of graphene, so are silicene and germanene, is the very small bandgap^[12], which limits their applications in high-end electronic nanodevices.

Different from the small bandgap materials, 2D monolayers of group-V elements are semiconductors with significant fundamental bandgaps, which have unique and promising electrical transport properties^[13], outstanding carrier mobilities^[14], topological transmission states, and negative Poisson’s ratio^[15], rendering them great application potentials in the next-generation functional nanodevices. 2D group-V monolayer materials include phosphorene, arsenene, antimonene and bismuthene, consist of phosphorus, arsenic, antimony and bismuth (see Fig. 1), respectively. Among these group-V elemental monolayers, the first explored and extensively studied one is phosphorene. Significantly, phosphorene has a tunable bandgap from 1.51 eV for a black phosphorus monolayer to 2 eV for a blue phosphorus monolayer, which would make it ideal for applications in nanoelectronic devices^[16-20]. Besides, two novel 2D wide-bandgap semiconductors, named as arsenene (monolayer As) and antimonene (monolayer Sb), were theoretically predicted^[21]. Both of them have significant bandgap with high stability, making them suitable for the optoelectronic devices. In addition, bismuthene, which is a monolayer of bismuth, has been found to be topologically nontrivial with a large spin-orbit-coupling gap^{[22, 23]}.

This review will sum up the latest progress on 2D group-V materials (phosphorene, arsenene, antimonene and bismuthene), especially the fabrication of group-V monolayers and their fundamental properties and applications. Finally, we will give a perspective on 2D group-V elemental semiconductors.

As a popular element in the Elemental Period Table, phosphorus has white, red, black and several other amorphous forms^[24-26]. Among all the amorphous forms of phosphorus, black phosphorus (BP) is the most studied and stable allotrope, which has a strong in-plane covalent bonding network with sp³-hybridized phosphorus atoms covalently bonded and weak van der Waals interactions between individual layers^[27], as shown in Fig. 2(a). The unique structural arrangement endows BP with amounts of novel properties, such as the tunable band gap from 0.3 eV for bulk BP to ~2 eV for monolayer BP^{[28, 29]} (Figs. 2(b) and 2(c)), moderate on/off ratio and high charge carrier mobility, which makes it ideal for the applications in field effect transistors^[30], optoelectronics, photodetectors^[31] and other devices. Since the discovery of layered 2D BP in 2014^[32], there have been more and more theoretical and experimental research on layered 2D BP. More recently, Baboukani et al. fabricated high-quality and orthorhombic crystal structured phosphorene through electrochemical exfoliation in deionized (DI) water, where BP power, stainless steel and DI water served as the raw material, electrode, and electrolyte respectively^[33]. The SEM and TEM images of BP nanosheets are shown in Figs. 2(d)–2(f).

Blue phosphorus, which was first introduced as “phosphorene”^[34], is 2D counterpart of black phosphorus. However, the structure of blue phosphorene has slightly flatter “zigzag” ridges rather than the deeper “armchair” ridges of black phosphorus^[35-38], as shown in Figs. 2(a) and 2(e). The first realization in experiment of large-scale 2D blue phosphorus was on a Au(111) substrate by molecular beam epitaxy (MBE) in 2016^[39]. The main preparation steps are as follows: First, phosphorus was evaporated from the crucible containing bulk black phosphorus at 260 °C, and then was deposited on a clean Au(111) surface for 3 min, which was heated at 230 °C. Finally, the substrate containing deposited phosphorus was annealed at 250 °C for 60 min. In this way, single layer phosphorus on Au(111) was successfully fabricated, as shown by STM image in Fig. 2(j). The high-resolution STM image and the corresponding line-profile (Fig. 2(j)) revealed a well-ordered hexagonal appearance of monolayer blue phosphorus. Interestingly, the product on the substrate, is blue phosphorus rather than black phosphorus. The main reason is probably due to the lattice mismatch between the black phosphorus and the substrate surface. This fabrication method of high-quality and large-scale single-layer blue phosphorus would be extended to many other 2D materials, which is hard to be achieved for large quantity by mechanical exfoliation. Moreover, Wang et al. reported the single vacancy (SV) defects could create mid-gap states within the fundamental bandgap of blue phosphorene^[40], as shown in Fig. 2(j) — perfect blue phosphorene with the bandgap of 1.94 eV and Fig. 2(k) — SV-(5|9) in blue phosphorene with the bandgap of 1.38 eV. In particular, SV-(5|9) in blue phosphorene shows considerable spin densities around the defect, as shown in Fig. 2(l), which paves a way for the application in blue phosphorene-based spintronics.

The metallic feature of multilayered arsenic is preserved until the thickness is decreased to a bilayer with a bandgap of 0.37eV, and to monolayer arsenene with a wide bandgap of 2.49 eV, predicted from density-function-theory (DFT) calculations in 2015^[21]. Theoretically, monolayer arsenene prefers a buckled honeycomb structure (Figs. 3(a) and 3(b)), instead of puckered and planar structure which are less favorable energetically^{[21, 43]}. In addition, external strain has been studied not only to turn indirect semiconductor arsenene into a direct band-gap semiconductor and semi-metal, as shown in Fig. 3(c), but also as a way to transform arsenene to a topological insulator, which has a quantum spin Hall state^[41]. Furthermore, Pizzi et al. in 2016 theoretically reported that arsenene was viewed as a promising 2D material for field effect transistors, which could apply in future nanoelectronics^[44].

So far, although the synthesis of buckled arsenene with monolayer thickness has not reported experimentally, a large amount of efforts has been paid to fabricate arsenene sheet in experiment. In 2017, Tsai et al. fabricated multilayer arsenene on InAs by the plasma-assisted method^[45], but its polycrystalline characteristics on the semiconductor substrate is not benefit for the study of its electrical properties.

Besides the plasma-assisted method, it has been recently reported that a high-yield arsenic nanosheets can be prepared by the method of aqueous shear exfoliation^[42]. Among the allotropes of arsenic, the most stable form is gray arsenic, which has the rhombohedral layered structure, making it very brittle and can be easily pulverized. The gray arsenic are used to fabricate arsenic nanosheets by the method of aqueous shear exfoliation. The fabrication process mainly contains three steps. The first step is the pre-treatment step, purifying the powders to remove oxides on bulk arsenic crystal surface. The starting materials (arsenic powder) were initially sonicated in aqueous surfactant sodium cholate (SC) in an ice bath. The suspensions were centrifuged, the powders were then dried at 60 °C in a vacuum oven. In the second step, the pre-treated powders were submitted in aqueous surfactant SC for shear dispersion and exfoliation by two types of kitchen-like blenders, i.e., a bottom shear mixer and a top shear mixer alternatively. In the process of shear exfoliation, use the bottom mixing in counter-clockwise and the top mixing in clockwise. In the third step, the suspensions were first centrifuged. Finally, the arsenic nanosheets were separated by aqueous washing and centrifugation at different rotation speed, and were dried at 60 °C in vacuum for follow-up studies.

Figs. 3(d) and 3(e) show the comparison between the starting materials and the final product of the shear exfoliation process. By using transmission electron microscope (TEM), the morphology of exfoliated arsenic nanosheets were shown in Fig. 3(f). The achieved arsenic nanosheets has visible wrinkled structure, which consists of few-layer materials. In addition, using the dynamic light scattering, the particles with diameters of arsenic nanosheets was investigated, which ranges between 100–200 nm dominantly.

Antimonene, which was predicted to have a bandgap of 2.28 eV^[21], has got more and more attention recently. This significant bandgap of antimonene implies its promising electronic and optoelectronics properties, which can be applied in optoelectronic devices. Recent studies have shown the fabrication of a few layers antimonene, but it is still difficult to grow high-quality monolayer antimonene^{[46, 47]}. Large-scale and high-quality buckled monolayer antimonene was achieved on a PdTe₂ substrate by MBE method^[48], which provides an effective way to achieve high-quality 2D materials.

The fabrication process of monolayer antimonene is illustrated in Fig. 4(a). The antimony atoms were evaporated from a Knudsen crucible and deposited onto a pre-cleaved PdTe₂ substrate kept at 400 K. Fig. 4(b) is a typical topographic image of as-resulted antimonene in a large-scale with a well-ordered LEED pattern inset, showing the high quality of the antimonene adlayer on the PdTe₂ substrate. Fig. 4(c) is a high-resolution STM image of antimonene adlayer, demonstrating a graphene-like honeycomb lattice of the antimonene. In addition, a buckled configuration of this honeycomb lattice can be seen in Fig. 4(d). Moreover, the periodicity of this honeycomb lattice is 4.13 ± 0.02 Å, which is quite close to the lattice constant of the PdTe₂ substrate (4.10 ± 0.02 Å). Therefore, this close match of these two lattices consistent with the LEED pattern showing (1 × 1) diffraction points (inset in Fig. 4(b)). The apparent height of antimonene is 2.8 Å (Fig. 4(e)), showing that the antimonene adlayer is a monolayer. A relaxed model with calculated electron localization function (ELF) value of 0.6 is shown in Fig. 4(f), illustrating the continuity of the adlayer antimonene.

Fig. 4(g) is a cross-section ELF pattern perpendicular to the monolayer antimonene, which is corresponded to the black line in Fig. 4(f). This image clearly shows two important EFL ranges: the ELF value of the Sb–Sb pair is larger than 0.8 and the ELF value of the strongest coupling region between the antimonene and substrate is less than 0.3. The comparison of these EFL ranges demonstrates that a strong interaction exists within the antimonene and a weak interaction exists between antimonene and PdTe₂ substrate.

The environmental stability of atomic-thickness 2D materials is of great importance. STM images of the sample before and after exposed in the air were shown in Figs. 4(h) and 4(i), which reveal that the antimonene island are clean without any impurity on the islands. Moreover, after annealing the sample at 380 K, the STM image of antimonene island (Fig. 4(j)) is quite similar to the initial state image. This evidence strongly proves the monolayer antimonene is quite inert to air. More recently, topological edge states have been found in buckled antimonene islands on top of antimonene monolayers, which is promising for future low-power electronic nanodevices and quantum computations^[49].

Compared to monolayer graphene with flat honeycomb lattice, group-V monolayer materials usually have a buckled structure. Interestingly, flat antimonene monolayer (FAM) is theoretically predicted and also realized by MBE growth, as reported by Shao et al. in 2018^[50]. FAM which inhibits hybridization between in-plane and out-of-plane orbitals, possess the quantum spin Hall effect, which would make them very suitable for applications in next-generation dissipationless electronic devices.

Theoretically, the s-orbital and p-orbital projected band structure of the buckled antimonene are calculated, as shown in Fig. 5(a). The hybridized bands around the Fermi level between the s, p_x, p_y, p_z orbitals of the Sb atom lead to the partially hybridized sp³ bands. As for the flat antimonene, there is no mutual hybridization between in-plane orbitals and out-of-plane ones as shown in Fig. 5(b). The in-plane orbitals lead to a hybridized sp² band. The out-of-plane orbitals lead to the Dirac cone of π bands resembling that of graphene around the Fermi level as denoted by the gray circles.

Experimentally, antimony atoms were deposited onto a pre-clean Ag(111) substrate kept at a relative low temperature of 353 K. A well-organized antimonene adlayer was formed on the substrate. Fig. 5(d) showed LEED pattern of the clean Ag(111) substrate with sharp (1 × 1) diffraction spots. Fig. 5(e) showed LEED pattern of the antimonene on the substrate, showing a ( × ) R30° superstructure with respect to the Ag(111) surface lattice.

A typical STM image of the antimonene adlayer is shown in Fig. 5(f), revealing a uniform film on the substrate. Fig. 5(g) is the close-up STM image of the sample, showing a clear honeycomb-like lattice. Measured along the profile line in Fig. 5(g), the period of 5.01 Å for the antimonene was obtained, which is consistent with the superstructure observed in the LEED patterns. Moreover, the height of the antimonene film is 2.2 Å, as shown in Fig. 5(h), which is lower than the buckled monolayer antimonene (2.8 Å)^[46], suggesting that the antimonene on the Ag(111) substrate has a lower buckling configuration. Based on first-principle calculations, the flat antimonene monolayer is the most energetically stable configuration on Ag(111) substrate, consistent with the experimental STM images.

Moreover, to determine whether Sb–Sb chemical bond exists in the flat configuration, the ELF of the antimonene monolayer on Ag(111) was carried out. Fig. 5(k) is the top view of the ELF for the optimized model of FAM, revealing the continuity of the antimonene. Along the black-line in Fig. 5(l) is the cross-sectional ELF, the different colors in different regions represent different interaction, suggesting strong chemical bond of the Sb-Sb pairs and weak Sb–Ag interaction. As for the configuration of this flat honeycomb lattice of FAM, it is sustained by the substrate. So, it is important to exfoliate these antimonene films from the substrate for its practical applications in the future, which need further exploring.

Given antimonene with the bandgap of 2.28 eV^[21], for bismuthene made of group-V elements, it is expected to have an even larger bandgap. Furthermore, as cousin of graphene, which was discovered with quantum spin Hall (QSH) effect, bismuthene has been particularly interested in search of the QSH state as it holds a similar honeycomb lattice with graphene^[51]. Monolayer bismuthene has been predicted to be topologically nontrivial^{[22, 23, 52]}. Therefore, realization of bismuthene as two-dimensional topological insulators is of great importance. Experimentally, a monolayer bismuthene with quantum spin Hall phase has been successfully realized on a SiC(0001) substrate by MBE method, as reported by Reis et al.^[53].

In Reis’s work, the use of SiC(0001) as the substrate is vital because this substrate is not only stabilizing the bismuthene as a quasi-2D topological insulator but also playing a great role in achieving a large gap. N-doped Si-terminated 4H-SiC substrate was first dry-etched in a hydro-based gas atmosphere at a temperature of ~1230 °C. And then the H-termination of the substrate was removed by a slow thermal H desorption at ~650 °C while immediately offer Bi atoms on the substrate. The substrate was then lowered to ~500 °C to condense Bi layer, then a × R30° superstructure of Bi atoms with a honeycomb lattice reconstructed. Figs. 6(a) and 6(b) denoted a comparation of LEED patterns before and after the formation of monolayer bismuthene. Fig. 6(a) showed an unreconstructed (1 × 1) surface resulting from the hydrogenation of the top-layer Si dangling bonds on the SiC(0001) substrate. Fig. 6(b) showed ( × ) R30° spots, indicating the formation of monolayer bismuthene on the surface after Bi deposition. Fig. 6(b) described schematically a bismuthene layer on the threefold-symmetric SiC(0001) substrate in ( × ) R30° commensurate registry. Obviously, this configuration of the bismuthene is planar instead of buckled, due to the strain effect of the SiC substrate. The STM image (Fig. 6(d)) of the sample shows the bismuthene film is smoothly covered on the substrate. The bismuthene flakes are extended of ~25-nm, limited by domain boundaries.

To establish the electronic structure of the bismuthene material, Reis et al. decomposed the band structure into σ-bond contributions came from Bi 6s, p_x, p_y orbitals and π-bond contribution from p_z orbitals. The orbitals at low energy are mainly σ bands, as shown in Fig. 6(e), as the hybridization of bismuthene and substrate is via π bonds. Considering the electronic structure around E_F, for the low energy bands without spin-orbital coupling (SOC), there is a Dirac-like band crossing at the K point, as shown in Fig. 6(f). However, when considering a strong atomic SOC, there is a large bandgap at the K point (Fig. 6(g)), and the Rashba term λ_R can lift the degeneracy of the valence-band and induces the spin-degenerate of the conduction band at K point (Fig. 5(h)). By using scanning tunneling spectroscopy (STS), the differential tunneling conductivity (dI/dV) curves show that while the interior of the island shows insulating gap feature, conductive states exists at the edges (Figs. 6(i) and 6(j)), which suggests that QSH exists in bismuthene/SiC. More excitingly, the bulk gap of bismuthene is up to 0.8 eV, which makes it a candidate for a high-temperature quantum spin Hall material. Very recently, the strong interaction electrons at the edge of the bismuthene on SiC(0001) was also reported to exhibit Luttinger liquid feature, providing a new platform to study one-dimensional electronic feature^[54].

There has been enormous progress on the study of group-V monolayers, especially on the preparation of new members and structural tuning of new materials. Phosphorene, which has amounts of excellent properties, such as tunable direct gap, high carrier mobility has been successfully fabricated on a Au(111) substrate by MBE. Arsenic nanosheets have been successfully prepared by aqueous shear exfoliation, which has visible wrinkled structure consisting of few-layer materials. As for antimonene, high quality buckled antimonene monolayer has been successfully grown on a PdTe₂ substrate, which represented excellent air stability in ambient environment. Moreover, flat antimonene monolayer has been also successfully fabricated on a Ag (111) substrate by MBE, which holding tunable topological properties for future applications. Regarding bismuthene, flat configuration as a two-dimensional topological insulator (2DTI) was experimentally obtained on a SiC (0001) substrate, with a large SOC gap, making it a candidate for a high-temperature quantum spin Hall material.

Undoubtedly, the experimental realization of group-V monolayers extend the scope of 2D atomically-thick structure and pave a way for the future application in spintronics. However, there are still some drawbacks in above-mentioned works on 2D monoelemental group-V monolayer materials. (1) Until now, the synthesis of buckled arsenene with monolayer thickness, which was predicted theoretically a wide bandgap, is not reported experimentally. (2) The flat antimonene monolayer films by MBE method are not electrically isolated from the metal substrate. (3) The excited findings are that the intriguing topological properties of group-V monolayers are reported experimentally in both antimonene and bismuthene monolayer, however, the low-energy-cost devices, based on the topological property of group-V monolayer, are still lacking, although there reported a lot of optoelectronic devices utilizing the semiconductor properties. So, there are many opportunities for research to exploit the structural controlling and potential applications in functional nanodevices. We hope that this review will inspire more attempts and findings in the group-V families, and we believe that related challenges can be solved in the near future.

Thanks the financial supporting from National Natural Science Foundation of China (Nos. 61971035, 61901038, 61725107), Beijing Natural Science Foundation (Nos. Z190006, 4192054), Strategic Priority Research Program of the Chinese Academy of Sciences (XDB30000000), and Beijing Institute of Technology Research Fund Program for Young Scholars (3050011181814).