Ferromagnetic ordering of monolayer vanadium dichalcogenides (VSe 2 and VS 2 ) has been predicted by density functional theory (DFT), and suggestive experimental evidence for magnetic ordering in VSe 2 monolayers has been reported. However, such ferromagnetic ordering would be in stark contradiction to the known paramagnetic nature of the bulk VSe 2 . Herein, we investigate the electronic structure of VSe 2 monolayers by angle-resolved photoemission spectroscopy (ARPES) and first-principles DFT. The ARPES measurements demonstrate the absence of spin-polarized bands for monolayers in close correspondence to nonmagnetic nature of the bulk VSe 2 . We demonstrate that the stabilization of the nonmagnetic state occurs due to the appearance of a charge density wave (CDW) state in VSe 2 monolayers. In contrast to well-established 4 × 4 × 3 periodicity of the CDW in bulk VSe 2 , we identify a √3 × √7 unit cell for VSe 2 monolayers from both scanning tunneling microscopy imaging and first-principles calculations. Moreover, DFT predicts that the √3 × √7 charge order state is energetically competitive with a ferromagnetic 1 × 1 state. This suggests that the experimentally observed CDW state is the nonmagnetic ground state of a perfect VSe 2 monolayer, consistent with the absence of spin-polarized bands in ARPES measurements. Therefore, monolayer VSe 2 is not an itinerant magnet.

Material growth by van der Waals epitaxy has the potential to isolate monolayer (ML) materials and synthesize ultrathin films not easily prepared by exfoliation or other growth methods. Here, the synthesis of the early transition metal (Ti, V, and Cr) tellurides by molecular beam epitaxy (MBE) in the mono-to fewlayer regime is investigated. The layered ditellurides of these materials are known for their intriguing quantum-and layer dependent-properties. Here we show by a combination of in situ sample characterization and comparison with computational predictions that ML ditellurides with octahedral 1T structure are readily grown, but for multilayers, the transition metal dichalcogenide (TMDC) formation competes with self-intercalated compounds. CrTe 2 , a TMDC that is known to be metastable in bulk and easily decomposes into intercalation compounds, has been synthesized successfully in the ML regime at low growth temperatures. At elevated growth temperatures or for multilayers, only the intercalation compound, equivalent to a bulk Cr 3 Te 4 , could be obtained. ML VTe 2 is more stable and can be synthesized at higher growth temperatures in the ML regime, but multilayers also convert to a bulk-equivalent V 3 Te 4 compound. TiTe 2 is the most stable of the TMDCs studied; nevertheless, a detailed analysis of multilayers also indicates the presence of intercalated metals. Computation suggests that the intercalation-induced distortion of the TMDC-layers is much reduced in Ti-telluride compared to V-, and Cr-telluride. This makes the identification of intercalated materials by scanning tunneling microscopy more challenging for Ti-telluride. The identification of self-intercalation compounds in MBE grown multilayer chalcogenides may explain observed lattice distortions in previously reported MBE grown early transition metal chalcogenides. On the other hand, these intercalation compounds in their ultrathin limit can be considered van der Waals materials in their own right. This class of materials is only accessible by direct growth methods but may be used as "building blocks" in MBE-grown van der Waals heterostructures. Controlling their growth is an important step for understanding and studying the properties of these materials.

Phase engineering has extensively been used to achieve metallization of two-dimensional (2D) semiconducting materials, as it should boost their catalytic properties or improve electrical contacts. In contrast, here we demonstrate compositional phase change by incorporation of excess metals into the crystal structure. We demonstrate post-synthesis restructuring of the semiconducting MoTe or MoSe host material by unexpected easy incorporation of excess Mo into their crystal planes, which causes local metallization. The amount of excess Mo can reach values as high as 10% in MoTe thus creating a significantly altered material compared to its parent structure. The incorporation mechanism is explained by density functional theory in terms of the energy difference of Mo atoms incorporated in the line phases as compared to Mo ad-clusters. Angle resolved photoemission spectroscopy reveals that the incorporated excess Mo induces band gap states up to the Fermi level causing its pinning at these electronic states. The incorporation of excess transition metals in MoTe and MoSe is not limited to molybdenum, but other transition metals can also diffuse into the lattice, as demonstrated experimentally by Ti deposition. The mechanism of incorporation of transition metals in MoSe and MoTe is revealed, which should help to address the challenges in synthesizing defect-free single layer materials by, for example, molecular beam epitaxy. The easy incorporation of metal atoms into the crystal also indicates that the previously assumed picture of a sharp metal/2D-material interface may not be correct, and at least for MoSe and MoTe, in-diffusion of metals from metal-contacts into the 2D material has to be considered. Most importantly though, the process of incorporation of transition metals with high concentrations into pristine 2D transition-metal dichalcogenides enables a pathway for their post-synthesis modifications and adding functionalities.

potentially to exploit them in spintronics applications. Such potential applications of magnetic vdW materials have recently sparked considerable interest in investigating magnetism of bulk ferromagnetic materials thinned to a single layer [1][2][3][4] or the emergence of ferromagnetism of paramagnetic bulk materials at the monolayer limit. [5,6] In addition, defect-and dopant-induced ferromagnetism has been predicted theoretically. [7][8][9][10][11][12] Recent experimental results suggest that 1T-2H phase boundaries in MoSe 2 , [13][14][15] edges in WS 2 or MoS 2 , [16][17][18][19] adsorbate induced defect states, [20] or substitutional doping of transition metals in 2D materials [21][22][23] can result in defect-or dopant-induced ferromagnetic ordering in these systems. Recently, long-range magnetic order has also been observed in both MoTe 2 and MoSe 2 , which has been suggested to be induced by intrinsic ferromagnetic defects, such as Mo antisites, that is, Mo atoms at chalcogen lattice sites. [24] This indicates that in these materials even dilute defects can cause long-range ferromagnetic ordering, making them promising diluted magnetic semiconductor (DMS) materials. Here, we investigate a new doping mechanism for 2H-MoTe 2 that enables altering the monolayer or surface layer of a MoTe 2 crystal with transition metal impurities. We have recently demonstrated that 2H-MoTe 2 and MoSe 2 can be modified by incorporation of transition metals into the host's interstitial site. [25] We have shown experimentally and by density functional theory (DFT) calculations that excess Mo atoms on MoTe 2 are energetically favored at interstitial sites as compared to adsorbed atoms at the surface. At elevated temperatures, these excess transition metal atoms are mobile and can undergo site-exchange with lattice Mo atoms. For high enough mobility (temperature), the interstitials rearrange into 1D Mo-rich crystal modifications, known as mirror twin grain boundaries. [26][27][28] These grain boundaries have shown no ferromagnetic properties. We demonstrate here, that this doping mechanism can also be expanded to other transition metals, in particular with the goal of inducing magnetism into MoTe 2 . Titanium or vanadium has been used as ferromagnetic dopants in diluted semiconductor systems [29,30] and thus, in this study we explore if V can be introduced into MoTe 2 interstitial Figure 4. Magnetization measurements of MoTe 2 with different V concentrations. a) M-H hysteresis loops taken at 10 K for MoTe 2 with 0.2, 0.3, and 0.8% of V coverage. The variation of the linear diamagnetic background is a consequence of different substrate thicknesses. b) Variation in magnetization saturation with the V concentration. c) Temperature dependences of H C and M S for the 0.8% V-doped sample. The error bars reflect the experimental uncertainty related to background noise. www.advancedsciencenews.com

Interlayer interactions in layered transition metal dichalcogenides are known to be important for describing their electronic properties. Here, we demonstrate that the absence of interlayer coupling in monolayer VTe2 also causes their structural modification from a distorted 1T′ structure in bulk and multilayer samples to a hexagonal 1T structure in the monolayer. X-ray photoemission spectroscopy indicates that this structural transition is associated with electron transfer from the vanadium d bands to the tellurium atoms for the monolayer. This charge transfer may reduce the in-plane d orbital hybridization and thus favor the undistorted 1T structure. Phonon-dispersion calculations show that, in contrast to the 1T′ structure, the 1T structure exhibits imaginary phonon modes that lead to a charge density wave (CDW) instability, which is also observed by low-temperature scanning tunneling microscopy as a 4 × 4 periodic lattice distortion. Thus, this work demonstrates a novel CDW material, whose properties are tuned by interlayer interactions.