The discovery of graphene^[1–3] opens a door of two-dimensional (2D) materials, bringing plenty of possibilities in material world. The extreme thickness of 2D materials limits the motion of electrons, resulting in quite different properties from their bulk counterpart. For example, graphene is transparent^[4] while graphite is not. Based on the novel properties, 2D materials hold great potentials in industrial applications.^[5–8] The exotic properties of 2D materials originate from diverse degrees of freedom, including the combinations of distinctive elements, different phases, and stacked orders. 2D materials could be composed of single-element or several elements, the former including graphene,^[9] silicene,^[10] germanene,^[11] borophene,^[12] stanene,^[13] the latter including transition metal chalcogenides,^[14,15] oxides,^[16,17] and so on. Phase also influences the physical properties dramatically. For instance, 2H-WSe₂ is a semiconductor^[18] while 1T’-WSe₂ demonstrates a metallic nature with enhanced electrocatalytic activity.^[19] Compared with bulk, 2D materials intrinsically feature the multiformity of stacking order. As was reported previously, CrI₃ displays layer-dependent magnetism, from ferromagnetic in monolayer, to antiferromagnetic in bilayer, and back to ferromagnetic in trilayer and bulk.^[20,21]

However, compared with 2D material with layered bulk counterpart, the research of 2D material without layered bulk counterpart is rarely reported. Recently, monolayer Cu₂Se and CuSe were experimentally synthesized and exhibited adjustable properties, whose bulk counterparts were nonlayered.^[22,23] Cu₂Se on bilayer graphene displayed a purely thermal structural phase transition at 147 K.^[22] Grown on Cu(111), CuSe was endowed with moiré patterns or nanopores due to lattice mismatch between CuSe and Cu(111) substrate.^[23,24] Besides, the property of CuSe is flexible, ranging from semiconductor to metal with Dirac nodal line fermions (DNLFs).^[24] These findings offer a route to explore new 2D materials with novel properties.

Here, we report the epitaxial growth of flat monolayer CuAs endowed with a honeycomb lattice on a Cu(111) substrate by molecular beam epitaxy (MBE). By combining characterizations of scanning tunneling microscopy (STM), low-energy electron diffraction (LEED), x-ray photoelectron spectroscopy (XPS), and first-principles calculations, we confirmed the atomic structures and superstructures of monolayer CuAs on the Cu(111) substrate. Angle-resolved photoemission spectroscopy (ARPES) measurements revealed the band structure of monolayer CuAs on Cu(111), which is further confirmed by DFT calculations. Further calculations show that the Fermi level of free-standing monolayer CuAs will be lift approximately 2.0 eV when the monolayer CuAs is put onto the Cu(111) substrate.

Monolayer CuAs films were grown by a straightforward arsenication of a Cu(111) substrate. Arsenic atoms were deposited onto a Cu(111) substrate kept at 470 K, and then the sample was annealed at 470 K to obtain an epitaxial monolayer CuAs film. A typical LEED pattern of the as-grown monolayer CuAs on Cu(111) substrate was obtained after growth, as shown in Fig. 1(a). The LEED pattern is composed of Cu(111) diffraction spots as well as unambiguous structures, which indicates a lattice constant of 0.443 nm for the monolayer CuAs. Figure 1(b) is a 200 nm × 200 nm STM image of the monolayer CuAs, showing overall fluctuations of the sample. No higher structure could be found except clear steps, indicating that the growth mode is layer by layer. In order to verify the chemical composition of the sample, we performed XPS measurements. Figure 1(c) is a typical As 3d spectrum, endowed with two prominent peaks at 41.06 eV (As 3 d_5/2) and 41.69 eV (As 3 d_3/2). The positions of the two As 3 d_5/2 peaks for our sample locate between those of elemental As (41.62 eV)^[30] and GaAs (40.9 eV),^[31] indicating that the chemical state of As in the sample is minus. In addition, a typical XPS spectrum for Cu shows a peak located at 932.4 eV, as shown in Fig. 1(d), which corresponds to Cu 2 p_3/2 and is slightly shifted from that of elemental Cu. The XPS results demonstrate that our sample is full crystallization and complete formation of monolayer CuAs on Cu(111) substrate.

Fig. 1.

Figure Option ViewDownloadNew Window

Fig. 1. STM, LEED, and XPS spectra of monolayer CuAs on a Cu(111) substrate. (a) LEED pattern of CuAs on Cu(111). White arrow denotes patterns from Cu(111), and red arrow denotes the diffraction spots of CuAs. (b) Large-area STM image of CuAs. (I = 0.05 nA, Vb = -1 V) (c) High-resolution XPS spectrum for As 3d. Two prominent peaks locate at 41.06 eV (As 3 d5/2) and 41.69 eV (As 3 d3/2). (d) XPS spectrum from the core level of Cu 2p. The peak at 932.4 eV is indicative of Cu 2p3/2.

To investigate the atomic structure of the monolayer CuAs, we performed a STM characterization. Figure 2(a) is a typical STM image of monolayer CuAs, showing a very flat terrace except some bright protrusions. Careful observations reveal that the protuberances are a chain structure. According to the experience of selenium atoms on metal substrate,^[32] we attribute the protrusions to extra arsenic atoms on monolayer CuAs. By constant height STM mode, atomic resolution STM image was achieved in Fig. 2(b), from which, uniform honeycomb lattice could be resolved. The lattice constant is measured to be 0.43 ± 0.02 nm, perfectly consistent with the LEED results.

Fig. 2.

Figure Option ViewDownloadNew Window

Fig. 2. Atomic structure of CuAs. (a) Close-up STM image of monolayer CuAs. The bright protuberances are extra As atoms. (I = 0.05 nA, Vb = –1 V). (b) Atomic resolution STM image of CuAs, clearly showing the honeycomb lattice. The black rhombus denotes a unit cell. (I = 5 nA, Vb = –20 mV). (c) Fully-relaxed atomic model of CuAs on Cu(111) substrate, depicting superstructure. The black rhombus denotes a unit cell. (d) Side view of CuAs, showing the planner structure.

In order to get a better understanding of the atomic structure of the CuAs layer, we perform DFT calculations and propose a planar honeycomb structure of CuAs as shown in Figs. 2(c) and 2(d). The arsenic atoms and copper atoms of CuAs are lying on the fcc sites of the Cu(111) surface and forming a honeycomb structure, which makes up a superstructure on the Cu(111) substrate. However, different from CuSe^[24] and ZnO,^[33] we do not observe moiré patterns on the CuAs/Cu(111) sample by STM and LEED, which indicates that the monolayer CuAs matches well with the Cu(111) substrate.

Electronic structure of monolayer CuAs was investigated by ARPES and DFT. Figure 3(a) displays ARPES data measured along the high symmetry direction M–Γ–K. We find that there is one hole-like band around the Fermi level. The second-derivative ARPES spectra of raw experimental band structures (Fig. 3(a)) are depicted in Fig. 3(b) to enhance the visibility of the hole-like bands. The second-derivative ARPES spectra agree well with the theoretical results. Here, two hole-like bands marked by red lines emerge, and the upper band crosses the Fermi level, indicating a metallic nature of monolayer CuAs on Cu(111) substrate. To attain more electronic band details, we performed DFT calculations. The calculated band structure is depicted in Fig. 3(c). A direct comparison between the ARPES data and the calculated band structure shows an excellent agreement.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. ARPES data and DFT calculated band structure of monolayer CuAs on Cu(111) substrate. (a) Energy–momentum dispersion along M–Γ–K direction in the hexagonal Brillouin zone. (b) Second-derivative spectrum of the raw ARPES data in (a). The overlaid lines mark the two bands. (c) First principle calculated band structure for monolayer CuAs on Cu(111).

Moreover, we investigate the intrinsic electronic structure of free-standing monolayer CuAs by DFT calculations. The top view and side view of the atomic structure of a monolayer CuAs are shown in Fig. 4(a). Monolayer CuAs has a flat structure with a honeycomb lattice. The calculated band structures of freestanding monolayer CuAs without and with the spin–orbital coupling (SOC) effect are shown in Figs. 4(b) and 4(c), respectively. In Fig. 4(b), four crossings of the three bands imply Dirac nodal line fermions (DNLFs).^[24] While considering SOC, a gap opens around the crossings. which can be attributed to the absence of the mirror reflection symmetry in the CuAs plane, which is similar with free-standing monolayer CuSe.

Fig. 4.

Figure Option ViewDownloadNew Window

Fig. 4. Calculated electronic structures of free-standing monolayer CuAs. (a) The top view and side view of atomic structure of a monolayer CuAs. The black rhomb marks the unit cell of monolayer CuAs. (b), (c) The band structures of a free-standing monolayer CuAs without and with SOC. (d), (e) Projected band structure of a free-standing monolayer CuAs with SOC. (d) The two bands marked in red are contributed mainly by in-plane orbitals (As px/py and Cu dxy/dx2 – y2). (e) The band marked in green is contributed mainly by out-of-plane orbitals (As pz and Cu dxz/dyz).

Compared with free-standing CuAs, the band structure of CuAs on Cu(111) changes mainly in two aspects. Firstly, when the monolayer CuAs is grown on Cu(111) substrate, the upward opening band along M–Γ–K direction of the Brillouin zone of free-standing CuAs disappears. To explain this phenomenon, the projected band structures for free-standing CuAs are calculated, as shown in Figs. 4(d) and 4(e). The two hole-like bands marked by red in Fig. 4(d) are contributed mainly by in-plane orbitals (As p_x/p_y and Cu d_xy/d_{x² – y²}), while the band marked by green in Fig. 4(e) is contributed mainly by out-of-plane orbitals (As p_z and Cu d_xz/d_yz). As a result, the out-of-plane orbitals of CuAs/Cu(111) are coupled by the substrate. Secondly, the Fermi level of monolayer CuAs on Cu(111) substrate is closer to the top of the hole-like bands, while the Fermi level will shift by about 2.0 eV in a freestanding CuAs. The Fermi level shift of monolayer CuAs on Cu(111) substrate should arise from substrate charge transfer.^[34] If more electrons could be transferred, the Fermi level will be lifted further and monolayer CuAs will transform into a semiconductor. Other possible methods to transfer electrons include surface doping of electron donor, like alkali metal atoms. Recently, it was reported that metallic monolayer honeycomb monochalcogenide MX (M = Cu, Ag; X = S, Se, Te) can be effectively tuned to semiconductor or topological insulator.^[35] We believe that the effect is also valid for CuAs since they have similar structure and composition, and exotic properties induced by other surface dopants and substrates are worth further investigations.

In summary, we have successfully synthesized high-quality, single-crystalline, monolayer CuAs by a direct arsenication of a Cu(111) substrate at 470 K. Characterizations by LEED, STM, XPS, and DFT calculations elucidated monolayer structure with honeycomb lattice. The ARPES measurements and their agreement with calculations revealed the metallic electronic structure of the monolayer CuAs. Additionally, first-principle calculated band structures of free-standing CuAs demonstrate that the charge transfer from substrate to monolayer CuAs modulates the Fermi level of monolayer CuAs. This work may enrich the family of 2D materials and provide a new candidate for nano-devices in the future.