Production of Tellurene Nanoribbons

ABSTRACT

The controllable synthesis of one‐dimensional (1D) nanostructures is foundational to nanotechnology but often relies on metal catalysts, introducing impurities and process incompatibility. Here, we demonstrate a catalyst‐free and broadly applicable strategy for the direct production of tellurene nanoribbons (TeNRs) via molecular beam epitaxy. The approach works on a wide range of exfoliated two‑dimensional (2D) materials and even on non‑layered substrates, although the degree of alignment and ribbon density are substrate dependent. Our approach leverages two synergistic growth paradigms. First, we employ van der Waals epitaxy on various exfoliated 2D materials, achieving in‐plane TeNRs. Second, we harness an edge‐attachment mechanism, where active step edges serve as anchoring points to initiate and direct the growth of freestanding TeNRs. Theoretical calculations reveal that this anisotropic growth is kinetically driven by a stark contrast between the stability of prismatic sidewalls and the dynamic reconstruction of the (0 0 0 1) growth front of Te. This approach provides precise kinetic control over morphology through substrate temperature and flux, yielding β‐phase TeNRs. Our work establishes an effective pathway for synthesizing elemental 1D nanostructures without catalysts, bridging the gap between surface‐guided and freestanding growth paradigms, with direct implications for integrating pure tellurium nanostructures into advanced electronics and optoelectronics.

A catalyst‐free molecular beam epitaxy strategy is reported for synthesizing tellurene nanoribbons (TeNRs). Leveraging substrate engineering, it combines van der Waals epitaxy on 2D materials with an edge‐docking mechanism at step edges. Anisotropic growth kinetics yields single‐crystalline β‐phase TeNRs, enabling clean integration into electronic and optoelectronic devices.

1. Introduction

The pursuit of one‐dimensional (1D) nanostructures [1], such as nanowire [2], nanoribbon [3], and nanotube [4], represents a cornerstone of nanotechnology, driven by their exceptional physical properties and potential for revolutionary applications in electronics, photonics, and quantum technologies [5]. Controllable synthesis of these nanostructures primarily relies on vapor deposition methods [6], which can be categorized into two distinct paradigms based on the growth interface and dimensionality.

The first paradigm is surface‐guided epitaxy [7]. One‐dimensional nanostructures grow laterally along a crystalline substrate surface [8], with their morphology, orientation, and crystallinity dictated by interfacial interactions, particularly along the engineered step edges [9, 10]. Moreover, in van der Waals (vdW) epitaxy [11], where epitaxy is no longer limited by the condition of lattice match, the strain can relax in a more effective manner [12], facilitating the growth of in‐plane nanostructures [13], and also providing the setting for heterostructures [14]. It has been proved that material with its own lattice constant can be grown by vdW epitaxy on a variety of layered material surfaces, such as transition metal dichalcogenides (TMDCs) [15]. The second paradigm aims for quasi‐freestanding growth, which is vertically oriented on the substrate [16], exemplified by the vapor–liquid–solid (VLS) [17] and vapor–solid–solid (VSS) mechanisms [18]. These principles have been successfully applied to fabricate a variety of 1D materials [6, 19], including III–V [20], II–VI compounds [21], oxides, and elemental semiconductors from Group IV (Si, Ge) [22]. In general, a catalyst, usually metal nanoparticles, is required to provide a nucleation point and to foster the gathering and the decomposition of precursors. However, the use of metal in growth is incompatible with complementary metal oxide semiconductor technology [23]. Although self‐assisted growth [24], using an element composing the grown material for VLS, has been used to assist the growth of GaP, GaAs, and InAs [25], achieving catalyst‐free, freestanding growth of 1D monoelemental nanostructures remains a formidable challenge [26, 27, 28], with confined growth inside carbon nanotubes being a notable but highly specialized exception [29, 30]. This critical gap underscores the significance of exploring inherently 1D elemental systems that could bypass the catalyst requirement altogether.

Tellurium (Te) emerges as a suited candidate to address this challenge [31, 32]. As the heaviest stable chalcogen, Te has gained significant attention in its two‐dimensional (2D) form—tellurene [33, 34, 35, 36]. As a p‐type elemental semiconductor, tellurene possesses high carrier mobility, excellent environmental stability, and remarkable optoelectronic properties, making it a compelling candidate for next‐generation electronics, photonics, and thermoelectric devices [37, 38]. Fundamentally, tellurium crystallizes in helical atomic chains that are stacked via vdW interactions, arising from hypervalent, three‐center four‐electron bonding configuration [39]. This quasi‐1D vdW‐bonded structure allows the crystal to be produced into 1D nanostructures without requiring external catalysts to break symmetry or define a growth axis [31, 40]. It offers an exceptional platform for the direct synthesis of high‐purity 1D elemental nanostructures in freestanding configuration.

Herein, we report the controlled production of tellurene nanoribbons (TeNRs) using molecular beam epitaxy (MBE), a technique offering ultra‐high vacuum (UHV) conditions and precise kinetic control, particularly for the phase and layer control by substrate engineering [41, 42, 43]. Our approach strategically employs two different growth mechanisms: (1) vdW epitaxy on various layered 2D material substrates, including tellurene, chalcogenides and graphene, which can be categorized into homoepitaxy and heteroepitaxy. This process is illustrated in Figure 1 (top panel), and further details are provided in the Experimental Section. (2) Edge‐attachment leading to a perpendicular growth of freestanding 1D nanoribbons (bottom panel). This dual‐pathway investigation allows us to decouple and elucidate the respective roles of lattice matching, interfacial energy modulation, and kinetic parameters in directing the anisotropic growth of TeNRs.

FIGURE 1.

Schematics of the MBE synthesis of TeNRs on diverse substrates. Solution‐grown 2D Te flakes are dispersed on SiO₂ substrates, followed by mechanical exfoliation to expose fresh terraces and edges; exfoliated layered 2D materials are prepared in parallel as alternative substrates. After loading into the UHV–MBE chamber, Te deposition yields two growth modes: (1) in‐plane vdW epitaxy on basal terraces and (2) edge‐anchored, quasi‐freestanding nanoribbon growth perpendicular to exfoliated edges via an edge‐triggered vapor–solid process.

2. Results and Discussions

2.1. Synthesis and Structural Characterization of TeNR on Te Flakes

Two‐dimensional Te flakes (micrometer size) were synthesized through the reduction of sodium tellurite by hydrazine hydrate in an alkaline solution at 160°C in the presence of polyvinylpyrrolidone (PVP) (see the Hydrothermal Synthesis of Tellurene Flakes subsection in the Experimental Section and Figure S1). The hydrothermal method in the presence of PVP leads to the formation of 1D Te nanostructures, on which the (1 0 $\overline{1}$0) facet is passivated at the initial stage. Extending the growth time enhances the growth along the 〈$\overline{1}$ 2 $\overline{1}$ 0〉 and 〈0 0 0 1〉 directions, thus leading to the formation of tellurene crystals [44]. The long axes (showing 1D characteristics) of these flakes are 〈0 0 0 1〉 oriented, and the lateral protruding regions grow along this direction, giving rise to a rectangular lattice of 2.2 Å and 6.0 Å (Figure 2a). The 2D Te flakes were transferred onto SiO₂/Si substrates to produce well‐dispersed thin films through a spin coating process. The resulting thin films were further mechanically exfoliated to produce a fresh surface. Figure 2b shows the scanning electron microscopy (SEM) image of a typical Te flake after mechanical exfoliation. The exfoliated 2D flakes show edge lengths ranging from 10 to 50 µm (Figure S2). These samples were loaded directly into the UHV chamber for MBE growth. For characterization, the as‐synthesized samples were transferred directly into SEM for imaging. Example micrographs are displayed in Figure 2c, and a gallery of SEM images is shown in Figure S3, which presents densely packed and unidirectionally aligned nanoribbons on the surface of the exfoliated Te flake. Large‐scale SEM images covering a whole Te flake are provided in Figure S4 showing the overall distribution of TeNRs on the flake. From these SEM images, we measured that the average width of TeNRs is ∼40 nm (plotted in Figure 2d). It is noted that the nanoribbons extend outside the flakes with the longer one twisted, which allows us to measure the thickness (<5 nm, Figure S4). Besides these TeNRs extending outside the terrace, we also observe the ones that anchor on the edges of Te flakes and protrude vertically along the $(\overline{1}2\overline{1}0)$ plane (Figure 2e). SEM images shown in Figure S5 present the well‐aligned TeNRs and isolated high‐aspect‐ratio ribbons with widths of less than 30 nm. From the SEM data of TeNRs grown on the exfoliated Te flakes, we measured the length and width of 963 individual ribbons. The scatter plot in Figure 2f shows the distribution, and a flake‑by‑flake breakdown is provided in Table S1. The average width is 35–42 nm, and the longest one is measured to be ∼1.2 µm. Histograms of the lengths, widths, and aspect ratios are shown in Figure 2f.

FIGURE 2.

Synthesis and characterization of TeNRs on solution‐grown tellurene flakes after exfoliation. (a) Atomic structure of a tellurene sheet (top and side views). (b) SEM image of a solution‐grown 2D tellurene flake after mechanical exfoliation. (c) SEM image of TeNRs grown by MBE, extending from the terrace of a tellurene flake. (d) Width distribution histogram for the terrace‐protruding TeNRs shown in (c). (e) TeNRs grown protruding outside along the flake edge, exhibiting perpendicular alignment and relatively uniform width. (f) Left: Scatter plot of length versus width for 963 TeNRs measured by SEM. See Table S1 for a detailed breakdown. Right: Corresponding histograms for the nanoribbons in Figure 2e, showing the distributions of length (top), width (center), and aspect ratio (bottom).

Control experiments confirm that the mechanical exfoliation step is essential for ribbon formation. This necessity arises because PVP molecules remain strongly adhered to the surface even after repeated washing cycles. Mechanical exfoliation effectively removes this polymer residue and exposes a fresh surface, thereby creating the essential conditions for the epitaxial growth of TeNRs. The exfoliated and non‐exfoliated Te flakes were further characterized by Raman spectroscopy and SEM experiments (Figure S6). The Raman features in the 1675–1687 cm⁻¹ range are clearly observed in non‐exfoliated flakes but are absent after mechanical exfoliation. MBE growth on non‐exfoliated Te flakes yields no TeNRs, whereas mechanically exfoliated flakes produce TeNRs both along the flake edges and on the flake surface, confirming that mechanical exfoliation removes PVP residues and exposes fresh surfaces/edges required for TeNR growth (Figure S6). Furthermore, as hinted by the Kossel crystal model [45], step edges rather than terraces are the energetically favorable sites for seed nucleation. The docking of an atomic seed at a step edge can lead to lowest‐energy configuration, which likely dictates the subsequent anisotropic growth. Evidence for the presence of such step edges is provided by atomic force microscopy (AFM) and SEM on the exfoliated Te flakes, as shown in Figure S2. These images clearly reveal terrace‑like features and height variations at the flake boundaries, confirming that the edges are not atomically flat but consist of steps that can serve as preferential nucleation sites. An atomic‐scale growth mechanism will be elaborated in Section 3.

2.2. TeNRs on Exfoliated 2D Materials

While TMDCs represent classic substrates for vdW epitaxy [46, 47], the underlying principle, utilizing atomically smooth and chemically inert surfaces, can be used to validate the universality of our growth method. Following mechanical exfoliation to obtain a clean surface, we performed Te deposition via MBE under identical conditions to investigate the potential for the universal vdW epitaxy of TeNRs on MoS₂ as an example. Figure 3a shows the SEM image of the grown Te on MoS₂. Differing from the thick rods and wires in previously reported physical vapor deposition (PVD) growth results [48], Te on the MoS₂ surface exhibits a needle‐like morphology. These TeNRs grow to 3–4‐µm lengths. The width of the most obtained TeNRs ranges from 50 to 100 nm, with almost half of the ribbons within less than 70 nm. These ribbons have a uniform thickness of ∼4 nm, as shown in the AFM image (Figure 3b and the corresponding line profile). In addition, these TeNRs are well oriented on the MoS₂ surface with a distribution at every 120° (statistical result from ∼100 ribbons in Figure S7). Considering the C_3v symmetry of the MoS₂ surface, the TeNRs are well aligned along specific crystallographic directions of MoS₂, suggesting a vdW epitaxial growth mode (see the theoretical analysis in Section 2.3, Growth Mechanism).

FIGURE 3.

MBE‐grown TeNRs on exfoliated MoS₂. (a) TeNRs grown on the MoS₂ terrace, exhibiting a threefold symmetric arrangement. (b) Width and height of in‐plane TeNRs on MoS₂ measured by AFM. (c) TeNRs grown attached to the edges of the MoS₂ flake. (d) Raman spectra collected from TeNRs grown on the terrace and at the edges of MoS₂.

Notably, nanoribbon growth is also not confined to the terraces but occurs along the edges of the exfoliated MoS₂ flakes. As shown in Figure 3c (and more images in Figure S8), the nanoribbons are firmly anchored to these edges, exhibiting a striking alignment where they tend to be parallel to each other and oriented perpendicular to the flake edge. This well‐defined geometry suggests a templating effect imposed by the substrate edges. Raman spectroscopy (Figure 3d) confirms the identity of these structures: the spectra from the ribbon‐covered areas (both on the terrace and edges) show a strong peak at 126  cm⁻¹ (A₁ mode), and a shoulder peak at 145  cm⁻¹ (E₂ mode), which are characteristic of the β‐phase tellurium, in full agreement with previous reports [46, 49].

We further demonstrate the broad applicability of our edge‐mediated growth strategy by successfully synthesizing TeNRs on a range of exfoliated 2D materials with diverse crystalline phases. Figure 4 presents a summary of SEM images of TeNRs grown on various 2D materials, confirming the general effectiveness of our approach. Consistent with the observations on MoS₂, both the terraces and edges of these 2D materials are decorated with TeNRs when using a substrate temperature of 373 K during MBE. Although the lattice constant changes abruptly at the interface, the crystal axes of the grown Te have been found to align well with those of the substrate.

FIGURE 4.

Growth of TeNRs on various exfoliated 2D materials. (a) SEM images (top row) of TeNRs grown on the indicated 2D materials via the MBE process, alongside their corresponding Raman spectra (bottom row), confirming the successful synthesis and phase identity of the tellurium nanostructures on each sample. (b) Key properties of the selected 2D substrates used in this study, including lattice constants and symmetry.

A comparative analysis of the growth outcomes reveals distinct morphological trends linked to the substrate identity (see lattice constants and atomic structures in Figure 4 and Figure S9). On WS₂ and SnS₂, edge‐anchored TeNRs exhibit exceptional length, extending to several micrometers. Growth on FePS₃ yields nanoribbons with superior orientation order and a relatively uniform distribution in both width (∼10 nm) and length (∼2 µm). Notably, at higher Te coverage, intersecting nanoribbons can merge to form quasi‐continuous films. Growth on CuInP₂S₆ under the same conditions is less prolific, yielding sparser nanoribbons. The edge‐anchored ribbons are much shorter, and nucleation does not occur on all available edges. Graphene presents a different case. While TeNRs on the terrace can reach ∼2 µm in length with a needle‐like profile, their orientation appears random with no clear lattice registry. Furthermore, growth directly on graphene edges is less favorable under T _sub = 373 K.

To underscore the universality of the edge‐mediated growth principle beyond vdW materials, we performed proof‐of‐concept experiments using non‐layered substrates with reactive edges. As shown in Figure S10a–b, TeNRs with micrometer length were successfully synthesized on the freshly fractured Si wafers. Similarly, patterned gold electrodes also served as effective anchoring templates for nanoribbon growth (Figure S10c–d). These results confirm that the strategy is not limited by the crystal structure of the substrate. Instead, it relies fundamentally on the presence of chemically active step edges, which can be provided by a wide array of materials, to initiate and guide the anisotropic growth of TeNRs.

2.3. Growth Mechanism

These growth systems are distinctly catalyst free. The nanoribbons form directly from the vapor phase via a vapor–solid (VS) mechanism [50]. The thermodynamic framework for such a process can be considered using the Gibbs’ phase rule (F = C −P + 2), where F, C, and P represent the degrees of freedom, the number of components, and the number of phases, respectively. In our case, with one component (Te, C = 1) and two phases (Te vapor and solid nanoribbon, P = 2), F = 1. This single degree of freedom is typically represented by temperature, implying that under a fixed vapor pressure, the system's equilibrium, and thus the driving force for nucleation and growth, is primarily governed by temperature. Consequently, the formation of nanoribbons from edge sites is a kinetically governed process, where the growth rate and morphology are highly sensitive to the substrate temperature [51].

To systematically evaluate the influence of growth parameters, we performed a series of control experiments on exfoliated Te by independently varying the substrate temperature, growth time, and Te source temperature (see Figures S13–S15). This configuration, reminiscent of the tip‐based growth seen in nanowire systems but without a foreign catalyst, was achieved only within a narrow temperature window (T _sub = 100°C) and required a high Te flux (T _source = 240°C). This suggests that a high supersaturation is necessary to drive the rapid, anisotropic nucleation at edge sites (see Figure S8). This aligns with the established understanding that precise control over substrate temperature and precursor flux is paramount in steering kinetically limited growth modes—a principle central to the synthesis of various 1D nanostructures [6].

To elucidate the kinetic origin of the anisotropic growth, we conducted a comprehensive density functional theory (DFT) study to identify the atomistic processes governing nanoribbon formation (see the Theoretical Calculations subsection in the Experimental Section). We first examined the stability of various low‐index surfaces on TeNRs. As shown in Figure 5(a) and Figure S11, the prismatic side facets, (1 0 $\overline{1}$0) and ($\overline{1}$2 $\overline{1}$0), remain unreconstructed and exhibit low surface energies of 0.55 and 0.67 eV/nm², respectively, confirming their high thermodynamic stability. In contrast, the (0 0 0 1) surfaces undergo reconstruction. Self‐passivation occurs through the rearrangement of Te atoms at the chain termini, effectively eliminating dangling bonds. For terminations involving an odd total number of atoms in two adjacent Te chains, the three most stable configurations are all reconstructed (Figure S11). For even‐numbered terminations, the reconstructed surfaces remain the most stable, followed by the unreconstructed ones (Figure S11). The surface energies of these six (0 0 0 1) configurations range from 1.63 to 2.88 eV/nm², substantially higher than those of the prismatic side facets. The distinct Te–Te bond configurations and stabilities of these facets provide critical insights into the origin of the anisotropic 1D growth.

FIGURE 5.

DFT‐calculated results. (a) Formation of energies of stable, exposed Te surfaces. (b) Atomic structures and adsorption energies for Te adatom on the (1 0 $\overline{1}$0) surface, the most stable (0 0 0 1)‐odd reconstructed surface, and the most stable (0 0 0 1)‐even reconstructed surface. Atomic structure of TeNR grown on the (1 0 $\overline{1}$ 0) surface along (c) the 〈0 0 0 1〉 direction and (d) the 〈$\overline{1}$2 $\overline{1}$ 0〉 direction. (e) Schematic illustration of the proposed kinetic growth mechanism of TeNRs. (1) Te atoms first adsorb on the step edges of the substrate. (2) The adsorbed Te atoms nucleate to form the initial chain structure, which serves as the embryonic state of the nanoribbon. (3) Subsequently deposited Te atoms preferentially attach to the termini of the existing chain. (4) The resulting odd–even terminated tip structure kinetically promotes the anisotropic 1D growth, leading to the elongation of the Te nanoribbon along the chain direction.

We next evaluated the kinetic behavior of different facets by calculating the adsorption of Te atoms on the tellurene plate, the dominant vapor‐phase species during MBE growth. Our calculations reveal a decisive contrast between the prismatic side facets and the reconstructed (0 0 0 1) surfaces. As shown in Figure 5b, the adsorption energy of a Te atom on the (1 0 $\overline{1}$0) surface is 0.63 eV, indicating weak adhesion and a kinetically hindered growth process. In contrast, adsorption on the low‐energy (0 0 0 1) reconstruction surfaces is markedly more favorable, with energies of 0.13 and −0.06 eV for the odd‐ and even‐numbered terminations, respectively. Furthermore, Te adsorption induces local structural interconversion between odd and even terminations. Te adsorption at an odd‐motif edge converts it to the even motif, while adsorption at an even motif reverses the process. This dynamic, reversible switching at the 〈0 0 0 1〉‐oriented growth front facilitates the continuous, stepwise advancement of the growth front that gradually extends along the 〈0 0 0 1〉 direction.

For the case of TeNR grown on exfoliated Te flakes, the preferential growth direction is further reinforced by the interaction between the ribbon and the substrate. We evaluated the binding stability and orientational preference of TeNRs on a Te (1 0 $\overline{1}$0) surface. The configuration aligned along the 〈0 0 0 1〉 direction (Figure 5c) exhibits a stronger binding (−0.294 eV/atom) than a misaligned one (−0.2 eV/atom, Figure 5d). This energetic preference for alignment is progressively amplified during growth, ultimately dictating the macroscopic orientation of the ribbons. The sequence of edge‑docking, chain nucleation, and anisotropic elongation is summarized schematically in Figure 5e, which highlights the kinetic pathway that leads to the formation of TeNRs.

Based on the theoretical calculations above, we further investigated the growth mechanism of TeNR on MoS₂. Our simulations reveal that Te atomic chains align orderly along the armchair direction of MoS₂ (Figure S12). This preferential orientation can be explained by the lattice matching relationship between MoS₂ (a = 3.16 Å) and Te (a = 4.52 Å, b = 5.95 Å along the 〈$\overline{1}$ 2 $\overline{1}$0〉 and 〈0 0 0 1〉 direction, respectively). Considering a configuration where a Te atom resides above a Mo atom, this site will repeat every 9.49 Å along the 〈$\overline{1}$2$\overline{1}$0> direction of Te, which corresponds to 3 × a(MoS₂) ≈ 2 × a(Te). Thus, the feasibility of vdW epitaxy is governed by the lattice match between the direction of Te chain extension and the underlying substrate. Further comparison of TeNRs on MoS₂ versus on the Te substrate reveals that the adsorption energy of a single TeNR chain on MoS₂ is only −0.036 eV/atom, substantially lower than that on the Te (1 0 $\overline{1}$ 0) surface. The equilibrium distance between the TeNR and the MoS₂ surface is 3.16 Å (Figure S12). We also plotted out the differential charge density, which reveals negligible interfacial charge transfer. Such weak interaction also enables the exfoliation and transfer of TeNRs from the MoS₂ surface. This mechanism similarly accounts for the uniform alignment of TeNRs on the FePS₃ substrate, where the lattice constants of Te (5.95 Å along the 〈0001〉 direction) and the FePS₃ (5.95 Å along the a‑axis) are perfectly matched, thereby promoting the unidirectional alignment of TeNRs. In contrast, TeNRs on graphene and CuInP₂S₆ exhibit low density, forming only sparse nanostructures. The lattice constant of graphene (2.46 Å) does not match an integer multiple of the TeNR lattice (5.95 Å), leading to random orientation. To further understand the poor growth behavior, we calculated the adsorption energy of a single Te chain on these two surfaces, yielding −0.197 eV/atom on graphene and −0.041 eV/atom on CuInP₂S₆ (Figure S12). The extremely weak adsorption on CuInP₂S₆ makes nucleation unfavorable, while the combination of moderate adsorption and severe lattice mismatch on graphene also suppresses well‑ordered growth.

Furthermore, freshly exfoliated edges of different 2D materials possess different densities of dangling bonds and chemical termination. These variations also affect the edge‑docking efficiency, and the nucleation barrier for Te atoms at the step edges. On CuInP₂S₆, which may contain mobile Cu⁺ ions [52], these Cu⁺ ions can readily react with the deposited Te atoms, forming copper telluride species rather than pure elemental TeNRs. Edge sites may be partially passivated or dynamically reconstructed, leading to sparse nucleation. On WS₂ and SnS₂, the zigzag edges [53] appear to be highly reactive, promoting long nanoribbon growth. The observation of TeNR growth on freshly fractured Si and Au electrodes confirms that the edge‑triggered mechanism is not limited to vdW‐layered substrates. On these non‑layered surfaces, the exposed step edges also provide anchoring sites that similarly promote the nucleation of TeNRs. Under our optimized kinetic conditions (T _sub = 100°C, high Te flux for Au), the ribbons still elongate perpendicularly to the edge direction, but the long‑range crystallographic alignment observed on single‑crystal 2D substrates (e.g., MoS₂) is absent because the non‑layered substrates are not well‑oriented single crystals. These results demonstrate that the essential requirement for the edge‑triggered growth is the presence of reactive step edges, regardless of the underlying crystal structure.

3. Discussion

By mastering the interplay between substrate design and growth dynamics, this work aims to establish a generalizable route for the atomically precise synthesis of 1D nanomaterials. Based on the experimental observations and theoretical calculations, we propose an edge‐triggered VS growth mechanism to explain the formation of freestanding TeNRs on the edges of various 2D crystals. The freshly exposed edges of exfoliated 2D crystals possess a significantly higher surface energy than their basal planes [54]. These edge sites act as preferential anchoring points with enhanced adsorption energy for Te atoms. This localized reduction in the nucleation energy barrier leads to the selective and dense nucleation of Te specifically along these edges, a process we term “edge‐docking.” The docked Te atoms utilize the periodic atomic structure of the edge as an initial template for alignment [55]. Subsequent growth proceeds via a highly anisotropic pathway through kinetic control. Adsorbed Te atoms experience a lower diffusion barrier for migration along the predefined, chain‐aligned edge than across the terraces, promoting longitudinal attachment. The main elementary processes involved in the TeNR growth, including adsorption–desorption kinetics and surface diffusion, can be quantitatively described in terms of the rate equation [56],

$$\frac{\partial c}{\partial t}=J-\frac{c}{\tau }+D\mathrm{\Delta }c$$

which incorporates an increase in the adatom concentration c due to the deposition of Te atoms with a flux J, a decrease in c due to the desorption of Te atoms with a residence time τ, and the supply of Te atoms due to surface diffusion. Impinging flux in the MBE technique is given by $J=P/\sqrt{2\pi m{k}_{\mathrm{B}}{T}_v}$ with P being the partial pressure of precursor, T_v the vapor temperature, m the particle mass, and k _B the Boltzmann constant. As the deposition flux of the gaseous precursor is typically low during MBE growth, the system could be viewed as a steady state, that is, ∂c/∂t = 0, which yields

$$-D\mathrm{\Delta }c+\frac{c}{\tau }=J.$$

It is evident that the steady‐state adatom concentration c is proportional to the deposition flux J. Moreover, the migration of Te atoms over the surface is largely promoted at elevated T _sub. Since the attachment rate scales linearly with the edge length, increased T _sub along with high deposition flux (and hence high adatom concentration) promote the generation of thermodynamically favorable 1D nanostructures [57]. The weak inter‐chain vdW interaction further suppresses lateral thickening. Consequently, the nanoribbon elongates predominantly along the direction perpendicular to the 2D edge from which it is seeded, growing outward into the vacuum with minimal interaction with the basal plane of the underlying substrate.

It is important to note that while our edge‑docking and vdW epitaxy strategy is effective on a wide variety of substrates, ranging from Te flakes and TMDCs to Si and Au, the growth behavior (e.g., alignment, ribbon length, and nucleation density) exhibits a clear dependence on the substrate. For instance, on graphene, the absence of commensurate lattice registry leads to random in‑plane orientation and inefficient edge‑anchored growth. Therefore, this approach shows a broad applicability of the underlying principles across diverse material systems, that is, catalyst‑free, edge‑triggered anisotropic growth under kinetic control. Finally, we acknowledge that one limitation of the current study is the inability to obtain direct lattice images of the TeNRs. This difficulty arises from the thick exfoliated growth substrate (not electron transparent) and the extreme fragility of the NRs, which precludes damage‑free transfer onto transmission electron microscope grids. Nonetheless, the well‑established Raman fingerprints of β‑Te, observed consistently across dozens of NRs on multiple substrates and under both growth modes, strongly support the phase identity and crystalline quality of the as‑grown nanostructures. Future scanning tunneling microscopy studies could directly resolve the atomic structure and edge configurations of these nanoribbons, a direct worth pursuing.

4. Experimental Section

4.1. Materials and Substrate Preparation

Unless otherwise specified, layered substrates used in this work were prepared by mechanical exfoliation. Bulk MoS₂ and WS₂ crystals were purchased from SUNANO Energy, and highly oriented pyrolytic graphite (HOPG) was purchased from Onway. Other layered crystals used for exfoliation, including SnS₂, FePS₃, and CuInP₂S₆, were grown by Dr. Lin Wang (Suzhou Institute of Nano‐Tech and Nano‐Bionics, Chinese Academy of Sciences) via chemical vapor transport (CVT).

For homoepitaxial growth, solution‐grown 2D Te flakes (tellurene) were used as precursor substrates. The Te flakes were dispersed onto SiO₂/Si substrates and subsequently exfoliated prior to MBE growth. For heteroepitaxial growth, mechanically exfoliated flakes of MoS₂, WS₂, SnS₂, FePS₃, CuInP₂S₆, and graphene were transferred onto 300 nm SiO₂/Si substrates using the adhesive exfoliation tape.

4.2. Hydrothermal Synthesis of Tellurene Flakes

Micrometer‐scale 2D Te flakes (tellurene) were synthesized by a hydrothermal reduction method. Na₂TeO₃ (0.10 g, 99.9%, Macklin) and polyvinylpyrrolidone (PVP, 0.11 g, M _w ≈ 5800, Macklin) were added to deionized water (20.8 mL) and stirred magnetically at room temperature until a homogeneous solution was obtained. The solution was transferred into a Teflon‐lined stainless‐steel autoclave, followed by the addition of the NaOH aqueous solution (7.0 mL, 0.096 mol L^{− 1}) and hydrazine hydrate (0.20 mL, 80%). The autoclave was sealed and maintained at 160°C for 23 h, and then naturally cooled to room temperature. The resulting silver‐gray product was collected by centrifugation (5000 rpm, 5 min) and washed with distilled water three times. A schematic illustration of the hydrothermal synthesis process is provided in Figure S1.

4.3. Mechanical Exfoliation

For both homoepitaxy and heteroepitaxy of TeNRs on layered crystals, substrates were mechanically exfoliated immediately before loading into the load‐lock of the MBE system to expose fresh terraces and step edges. In the case of hydrothermally synthesized Te flakes, exfoliation also removed surface polymer residues, which facilitated epitaxial nucleation and TeNR growth. Control experiments on non‐exfoliated Te flakes did not yield TeNRs, indicating that freshly exposed surfaces/edges are essential for nanoribbon formation. The exfoliated Te flakes typically exhibited lateral dimensions of approximately 10–50 µm.

4.4. MBE Growth of TeNRs

TeNRs were grown in a UHV–MBE (Fermion Technology) system using elemental Te as the evaporation source. Exfoliated materials (Te flakes and 2D layered crystals on SiO₂/Si) were directly loaded into the UHV chamber for growth.

By tuning growth kinetics, two growth modes were obtained: (1) in‐plane vdW epitaxy on basal terraces and (2) edge‐docking quasi‐freestanding nanoribbon growth initiated from exfoliated edges.

To systematically evaluate the influence of growth parameters, we performed a series of control experiments on exfoliated Te by independently varying the substrate temperature, growth time, and Te source temperature (see Figures S13–S15). The formation of well‑defined, high‑aspect‑ratio TeNRs occurs only within a narrow kinetic window: T _sub = 100°C, T _source = 240°C, and growth time of ∼10 min. The edge‐docking growth mode was observed only within such a narrow process window under relatively high Te flux. In our system, temperature‐ and time‐dependent growth experiments were typically performed at T _source = 240°C, with substrate temperatures of 100°C, 125°C, and 150°C, and growth durations of 5, 10, and 15 min. These observations confirm that the growth mode is strongly governed by the interplay of substrate temperature, Te flux, and deposition time.

4.5. Growth on Non‐layered Substrates

To assess the feasibility of the edge‐docking growth mechanism beyond layered vdW substrates, we conducted proof‐of‐concept experiments on non‐layered substrates featuring reactive edges, including freshly fractured Si surfaces (PVD) and thermally evaporated Au electrodes (MBE). Micrometer‐scale TeNRs were observed at these reactive edges, supporting an edge‐triggered growth mechanism that is not restricted to layered crystal substrates.

4.6. Scanning Electron Microscopy

Morphology characterization was performed primarily by SEM. SEM was used to image exfoliated Te flakes and TeNRs on a variety of substrates. The dimensional statistics reported in the main text were obtained from SEM measurements of 963 individual nanoribbons. SEM imaging was carried out using a Thermo Fisher Quattro ESEM operated at 10 kV.

4.7. Raman Spectroscopy

Raman spectroscopy was used to identify the Te phase and to compare terrace‐ and edge‐grown structures on layered substrates. Characteristic β‐Te Raman modes (A₁ and E₂) were observed in TeNR‐covered regions. Raman spectra were acquired on a Renishaw inVia Qontor system using a 532‐nm excitation laser, a 100× objective, and a laser power of 100 µW.

4.8. Atomic Force Microscopy

AFM was used to measure the thickness and width of TeNRs on MoS₂ terraces. AFM measurements were performed at Shanghai Jiao Tong University using a commercial AFM instrument. PeakForce Tapping mode from the Bruker Dimension ICON acquired AFM images of the TeNRs.

4.9. Theoretical Calculations

We explored the low‐energy (0 0 0 1) Te surfaces through an unbiased global search using the Crystal structure AnaLYsis by Particle Swarm Optimization (CALYPSO) software package [58, 59] in combination with first‐principles calculations. Throughout the search process, 60% of the 20 candidate structures in each generation were generated by the Particle Swarm Optimization algorithm, while the remaining 40% were randomly constructed.

The Vienna Ab initio Simulation Package (VASP) [60, 61] based on DFT was employed to optimize all structures in this study. The interaction between valence electrons and ion cores was described using the projector augmented‐wave method [61, 62], while exchange–correlation interactions were treated using the Perdew–Burke–Ernzerhof functional [63, 64]. The DFT‐D3 method [65] was adopted to account for the vdW interactions between Te chains and MoS₂. A plane‐wave basis set with a cutoff energy of 500 eV was used in all calculations. All structures were fully relaxed, with the energy and force convergence criteria set to 10⁻⁵ eV and 10⁻² eV Å⁻¹, respectively.

Surface energies (E _f) were calculated using the formula E _f = (E _slab − E _bulk) / 2A, where E _bulk represents the energy of bulk Te, E _slab denotes the energy of the fully optimized Te slab, and A is the area of the exposed surface.

5. Conclusion

This study established a catalyst‐free route for the synthesis of 1D TeNRs through MBE. By leveraging two synergistic growth paradigms, vdW epitaxy on inert basal planes and edge‐triggered vapor–solid growth at reactive step edges, we achieve deterministic control over the orientation, morphology, and dimensionality of the resulting 1D TeNRs. Comprehensive DFT calculations unveil the kinetic origin of this anisotropic growth, attributing it to a pronounced asymmetry between passivated prismatic facets and a dynamically reconstructing (0 0 0 1) growth front. The demonstrated compatibility of our strategy with a diversity of substrates, spanning various lattice symmetries, and chemical compositions underscores its generalizability beyond traditional lattice‐matching constraints. This work bridges the gap between surface‐guided and freestanding growth modes, providing a versatile platform for the direct integration of high‐purity, functional 1D building blocks into advanced electronic and optoelectronic architectures, while opening new avenues for the synthesis of other elemental nanostructures without the need of foreign catalysts.

Conflicts of Interest

The authors declare no conflict of interest.

Supporting Information

Schematic of hydrothermal synthesis of tellurene flakes; SEM and AFM characterization of Te flakes and surface terraces; SEM images of MBE‐grown TeNRs on Te flakes and exfoliated 2D substrates; Raman characterization and control MBE growth experiments before and after mechanical exfoliation; edge‐docking growth of TeNRs on MoS₂; atomic structures and lattice parameters of layered 2D materials; proof‐of‐concept TeNR growth on Si and Au edges; DFT‐calculated Te facet structures, surface energies, and adsorption configurations of Te chains on representative 2D substrates; morphology evolution under different growth parameters; and statistical summary of 963 TeNRs.

Supporting information

Supporting File: smtd70841‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.