1 Introduction
In recent years, topological materials [
1−
6] have emerged as a research frontier due to their unique electronic structures and potential applications in novel nanoelectronic devices [
7−
11]. Among diverse topological materials, bismuth halides Bi
4X
4 (X = Br, I) exhibit their rich topological phases [
12−
17] and diverse physical properties [
18−
22]. Theoretical and experimental studies confirm monolayer Bi
4Br
4 is a quantum spin Hall insulator possessing a large bandgap (~0.2 eV) [
23−
26], while multilayers allow multi-channel transport via weak interlayer coupling [
27,
28]. Notably, bulk crystals display phase-dependent topology [
12,
13,
26,
29,
30]: the
α-phase is a higher-order topological insulator with topologically protected hinge states [
26,
31−
34], while the
β-phase is a weak topological insulator [
26,
35−
38], together providing a versatile platform for exploring low-dimensional topological quantum effects [
33,
39−
43]. Structurally, Bi
4X
4 consists of van der Waals-assembled one-dimensional (1D) molecular chains [
26,
29,
44,
45], forming a chain-like crystal architecture with pronounced anisotropy [
19,
22,
33,
41], exhibiting distinctly different physical properties along the chain (
b-axis) versus perpendicular directions.
A key feature of this material is its topologically protected 1D edge states [
24,
28,
31,
32,
46], which enable dissipationless transport [
33,
41,
43,
47,
48], and the pronounced electron confinement in its quasi-1D structure further provides an ideal platform for investigating strong correlation phenomena like Luttinger liquid behavior [
49,
50]. However, current research relies primarily on bulk crystals and mechanical exfoliation. The inherent quasi-1D chain-like structure typically yields flakes with limited lateral dimensions [
22,
33], making it difficult to obtain large-area thin layers. On the other hand, existing direct synthesis methods generally suffer from the problem of producing nanostructures with disordered [
51] or multi-directional orientations [
24,
52,
53], making the growth of well-aligned arrays a persistent challenge in the field. These limitations prohibit systematic studies of intrinsic anisotropic properties and device integration of topological edge states. To bridge this gap, it is essential to develop directed synthesis techniques for producing aligned, atomically precise “quantum wires”. Such controlled architectures would enable high-performance interconnects in integrated circuits [
54,
55], reducing resistive heating and energy loss. Moreover, they represent a critical prerequisite for coupling nanowires to superconducting electrodes in Majorana-based quantum computing hardware [
56,
57].
In this work, WTe2 is selected as an epitaxial substrate to achieve oriented growth of Bi4Br4 nanoribbons. The anisotropic surface of Td-phase WTe2 effectively guides 1D Bi4Br4 to align along specific crystallographic directions, overcoming the random orientation typical of previous methods. Through synergistic optimization of key parameters including growth temperature, source rate ratio and annealing process, we established the optimal growth window for high-quality nanoribbons. The as-prepared nanoribbons possess well-defined morphology, sharp edges, uniform b-axis alignment and thickness continuously tunable from 2 to 9 nm (≈ 2–9 layers). The oriented Bi4Br4 not only provide a unique system for exploring anisotropic physics and correlated topological effects in the 1D limit but also establishes a critical material foundation for constructing functional devices.
2 Experiential section
2.1 Sample preparation and growth procedure
The WTe2 substrate was mounted onto a tantalum plate using silver conductive paste (EPO-TEK H20E, USA), and the assembly was secured onto a flag-style sample holder via spot-welded Ta wires, see Fig. S1 in the Electronic Supplementary Materials (ESM). To obtain an atomically flat and clean surface, the WTe2 substrate was mechanically exfoliated using 3M Scotch tape, effectively removing surface oxides and adsorbed contaminants. The exfoliated sample was promptly transferred into the load-lock chamber. After reaching a base pressure of 5 × 10−5 Pa, it was transferred to the MBE growth chamber, where the base pressure was maintained at approximately 1 × 10−7 Pa. The substrate was subsequently annealed at 200 °C for 2 hours under ultra-high vacuum to further desorb residual surface adsorbates, creating an ideal interface for subsequent high-quality van der Waals epitaxy.
Following substrate pretreatment, we set the system to the target growth temperature to commence thin-film deposition. High-purity Bi (99.999%) and BiBr
3 (99.9%) were loaded into separate high-temperature effusion cells, allowing independent control of their deposition rates through temperature regulation. During deposition, the Bi deposition rate was monitored in real-time using a quartz crystal monitor. The rate was first stabilized at 0.25 BL min
−1, with 1 BL corresponding to the single-step height of Bi (111) [
58]. The shutter was then opened to begin deposition, and the process duration was varied from 5 to 30 minutes depending on the desired thickness. Subsequently, the pre-deposited Bi layer was annealed at the same temperature, followed by exposure to BiBr
3 vapor. This step enabled the reaction between Bi and BiBr
3 to form stoichiometrically controlled Bi
4Br
4 thin films, ultimately yielding a Bi
4Br
4/ WTe
2 van der Waals heterostructure.
2.2 Materials characterization
The structural and chemical properties of the synthesized materials were characterized using a suite of analytical techniques. Surface morphology and structural features were examined by atomic force microscopy (AFM) on a Bruker Multimode 8 system and a Bruker Dimension XR system. Elemental composition and chemical states were analyzed by X-ray photoelectron spectroscopy (XPS) using a PHI QUANTERA-II SXM instrument with a monochromatic Al Kα X-ray source (1486.6 eV). Crystal structure was assessed via X-ray diffraction (XRD) on a Bruker D2 Phaser diffractometer with Cu-Kα radiation (λ = 1.54184 Å). Quantitative elemental analysis was performed using a JEOL JSM-IT700HR scanning electron microscope (SEM) equipped with an energy-dispersive X-ray spectroscopy (EDS) detector. Raman spectroscopy measurements were conducted on a WITec alpha 300 system with a 532 nm laser source and a 50× objective to probe lattice vibrational modes. The lattice structure of the nanoribbons and their epitaxial relationship with the WTe2 substrate were characterized by scanning tunneling microscopy (STM, CreaTec LT-STM system) and transmission electron microscopy (TEM, Thermo Scientific Talos F200X system). Cross-sectional lamellae for TEM observation were prepared using a focused ion beam (FIB, Thermo Scientific Helios 5 UC system).
3 Results and discussion
Bismuth bromide (Bi
4Br
4), a member of the bismuth halide Bi
4X
4 family, has recently emerged as a promising topological quantum material [
13,
25,
31−
33,
43], attracting broad interest due to its distinctive electronic structure and potential applications. Bulk Bi
4Br
4 crystallizes in a monoclinic structure with the space group C2/m [
32]. The unit cell consists of 1D molecular chains extending along the
b-axis, which align parallel along the
a-direction, with each layer containing two such chains [Fig. 1(a)]. The stacking sequence between layers determines the polytype; to date, only the
α-phase has been experimentally confirmed in bulk crystals [
13]. In this phase, layers adopt an AB stacking sequence along the
c-axis, accompanied by a half-unit-cell shift along the
b-axis and a 180° rotation between adjacent layers [
29]. Within each chain, Bi atoms form a zigzag backbone via strong covalent bonding, while Br atoms terminally bond to outer Bi sites, forming near-vertical configurations with bond angles
α ≈ 91° and
β ≈ 92°. The corresponding lattice parameters are
a = 13.064 Å,
b = 4.338 Å, and
c = 20.061 Å [
13]. The pronounced contrast between strong intrachain covalent bonds and weak interchain/interlayer van der Waals interactions facilitates cleavage along the
b-axis or the epitaxial growth of high-quality nanoribbons without dangling bonds [
13,
24,
59], making Bi
4Br
4 an ideal system for constructing topologically robust 1D electronic channels.
Structurally analogous, tungsten ditelluride (WTe
2) also exhibits strongly anisotropic quasi-1D character. As a representative layered transition metal dichalcogenide (TMDC), it adopts a thermodynamically stable orthorhombic Td phase (space group Pnm2
1, No. 31) under ambient conditions [
60]. Its structure can be viewed as a strongly distorted variant of the ideal 1T phase, composed of distorted Te–W–Te sandwich layers stacked along the
c-axis via van der Waals interactions [Fig. 1(a)]. Unlike high-symmetry 2H-phase TMDCs, both W and Te atoms in WTe
2 undergo dimerization displacements, forming 1D zigzag chains along the
a-axis and leading to marked structural anisotropy. The lattice constants are
a = 3.483 Å,
b = 6.278 Å, and
c = 14.054 Å. This low-symmetry structure breaks inversion symmetry, which is considered the structural origin of its extraordinary properties [
61], such as extremely large unsaturated magnetoresistance [
62] and type-II Weyl semimetal states [
63]. It is noteworthy that both WTe
2 and Bi
4Br
4 exhibit strong uniaxial anisotropy along their respective axes (WTe
2:
a-axis; Bi
4Br
4:
b-axis). This structural commonality makes WTe
2 an ideal substrate for guiding the oriented epitaxial growth of Bi
4Br
4. Thus, such a TMDC–topological insulator van der Waals heterostructure provides a distinct platform to explore novel interfacial quantum effects [
64,
65], such as charge transfer or enhanced spin–orbit coupling.
Here, we demonstrate the oriented epitaxy of Bi4Br4 nanoribbons on WTe2 substrates using molecular beam epitaxy (MBE). AFM [Fig. 1(b)] reveals that the nanoribbons with regular shapes and sharp edges extend along their intrinsic b-axis, aligning strictly parallel to the a-axis of the WTe2 substrate, demonstrating excellent epitaxial orientation uniformity. Cross-sectional HRTEM, cut perpendicular to the nanoribbon long axis, further confirms this epitaxial relationship (Fig. S2): the WTe2 atomic arrangement corresponds to its bc-plane, showing that the nanoribbons grow oriented along the substrate’s a-axis. Moreover, a sharp and well-defined interface is observed, where lattice fringes on both sides terminate at the interface rather than continuing across it, directly evidencing van der Waals epitaxy. Severe lattice damage and thermal decomposition induced by FIB thermal effects rendered the nanoribbon’s intrinsic atomic-resolution features unobservable. HRTEM-EDS mapping (Fig. S3) reveals Bi/Br in the upper film and W/Te in the underlying substrate. Unlike uniform Bi, Br is scarce and more diffuse due to heat-induced Bi–Br bond breaking and Br volatilization during FIB preparation.
These nanoribbons have an average width of ~15 nm and lengths predominantly in the 70–130 nm range. Cross-sectional height analysis [Fig. 1(c), along the red line in the inset] reveals a height distribution of 5–6 nm. Given that the monolayer thickness of
α-Bi
4Br
4 is ~0.96 nm [
25,
33], this height corresponds to stacks of approximately 5–6 molecular layers. Importantly, in multilayer Bi
4Br
4, weak interlayer coupling facilitates topological edge state decoupling at the terraced edges formed during layer-island growth and maintains independent boundary states across different layers [
27], thereby offering an ideal platform for multi-channel, low-dissipation electronic transport.
To further confirm the chemical composition and bonding states of the nanoribbons, XPS analysis was performed. The survey scan [Fig. 1(d)] shows characteristic signals originating only from the substrate (W, Te) and the target film (Bi, Br), apart from the C 1s and O 1s peaks attributed to ambient exposure. No other impurity elements were detected, indicating high chemical purity of the sample. High-resolution XPS spectra of the Bi 4f and Br 3d regions are presented in Figs. 1(e) and (f), respectively, providing further insight into the chemical environment. In the Bi 4f region [Fig. 1(e)], a doublet appears at 157.4 eV and 162.7 eV, corresponding to the 4f
7/2 and 4f
5/2 levels in a Bi–Bi bonding configuration. Another doublet observed at higher binding energies of 159.4 eV and 164.8 eV is assigned to the corresponding orbitals in a Bi–Br bonding environment. Simultaneously, the Br 3d spectrum [Fig. 1(f)] exhibits a well-defined doublet at 68.7 eV (3d
5/2) and 69.7 eV (3d
3/2), further confirming the formation of Br–Bi covalent bonds. All observed binding energies are in excellent agreement with literature values for Bi
4Br
4 [
24,
52]. Semi-quantitative XPS analysis (shown in Table S1 in the ESM) gives a Bi:Br atomic ratio of 20.09:26.31, which is close to the 1:1 stoichiometry of Bi
4Br
4, supporting the successful preparation of high-purity quasi-1D Bi
4Br
4 nanoribbons.
Raman and XRD further characterized the molecular vibrational modes and crystal structure of the heterostructure. The Raman spectrum [Fig. 2(a)] displays distinct characteristic peaks at 81, 114, 134, 163, and 212 cm
−1, corresponding to the
2A
1,
3A
1,
5A
1,
7A
1, and
9A
1 modes of Td-WTe
2, respectively [
66]. However, the strong overlap of the WTe
2 peak at 114 cm
−1 with the Bi
4Br
4 primary peak at 115 cm
−1 [
33] severely interferes with the clear detection of the film’s Raman signal. Raman measurements on Bi
4Br
4 bulk crystals (Fig. S4) reveal that the intrinsic peaks (115 cm
−1 main peak, with two weak peaks at 102 and 125 cm
−1) are present only at a low laser power of 0.3 mW. Higher powers induce thermal damage, causing additional peaks or complete disappearance of the intrinsic signals, confirming the high thermal sensitivity of Bi
4Br
4. On the WTe
2 substrate, a clear Raman spectrum requires 2–3 mW — a power level that would thermally damage Bi
4Br
4 and eliminate its intrinsic signals. To avoid both peak overlap and thermal damage, the Bi
4Br
4 film grown on a silicon substrate was measured at 0.3 mW (Fig. S5), showing a clear characteristic peak at 116 cm
−1.
Benefiting from the highly uniform in-plane orientation of the nanoribbons, we increased their areal density and grain size to enhance the XRD signal and applied a logarithmic scale (log scale) on the vertical axis to amplify weak signals, while simultaneously minimizing interference from the strong diffraction peaks of the substrate. As shown in Fig. 2(b), we observed diffraction peaks labeled in orange at 2
θ = 28°, 37°, 48°, and 59°, which correspond to the (00l) family of planes of Bi
4Br
4 [
19,
41]. Peaks originating from the WTe
2 substrate and the tantalum plate are labeled in black and gray, respectively.
To further confirm the crystal structure, STM images parallel to the nanoribbon long side reveal periodic lattice fringes [Fig. S6(g)] with a 0.4 nm spacing, matching the b-axis constant (4.34 Å) of monoclinic Bi4Br4, confirming that the long side is the b-axis and consistent with Bi4Br4’s anisotropic chain-like structure. The 0.6 nm fringe length is about half the a-axis constant (13.06 Å), and the lattice image matches the ab-plane atomic model [Fig. S6(h)], verifying the Bi4Br4 lattice. The WTe2 lattice fringe pattern [Fig. S6(a)] shows ab-plane characterized by chain-like structures along the a-axis, with a 0.6 nm fringe spacing matching its b-axis constant (6.28 Å). This demonstrates that the nanoribbon b-axis aligns with the WTe2 a-axis — direct evidence for substrate-lattice-guided oriented growth. Figures S6(b)–(d) show nanoribbons with regular shapes, sharp edges, and flat surfaces, reflecting high crystalline quality. Thus, the STM results confirm the epitaxial relationship and demonstrate at the lattice scale that the nanoribbons are Bi4Br4, not other bismuth bromides.
Meanwhile, two annealing treatments reveal significant thermal instability of the Bi4Br4 nanoribbons: the elongated nanoribbons gradually shorten with annealing [Figs. S6(e, f)], and their surfaces even exhibit segmentation, indicating signs of fracture [Figs. S6(i)–(k)]; a few have transformed into spherical and triangular Bi grains, as elevated temperatures induce thermal decomposition, consistent with our HRTEM observations (Figs. S2 and S3), high-temperature growth and annealing studies (Figs. 3 and 4).
Further evidence for the formation of the heterostructure comes from EDS analysis. The EDS spectrum [Fig. 2(c)] reveals a Bi:Br atomic ratio of 0.40:0.55 (approximately 1:1) and a W:Te ratio of 28.19:56.38 (1:2), both matching their expected stoichiometries and agreeing with the XPS quantitative analysis results in Table S1. Other than C, O, and Al signals attributable to the environment and instrument, no impurity-related peaks are observed. The EDS elemental mapping in Fig. 2(d) clearly reveals that Bi, Br, W, and Te are uniformly distributed on the heterostructure surface. These EDS results, together with the morphological, structural, and chemical characterization presented above, demonstrate the successful fabrication of the Bi4Br4/WTe2 van der Waals heterostructure.
We developed a two-step epitaxial growth method to synthesize Bi4Br4 nanoribbons on WTe2 substrates. The process consists of depositing a Bi pre-layer, followed by the introduction of BiBr3 vapor as a bromine source, enabling Bi4Br4 crystallization within an appropriate temperature window. The key to this method depends on precise substrate temperature control and independent rate monitoring to ensure an exact source ratio — both of which are essential for achieving high-quality morphology and crystallinity. We systematically investigated the influence of substrate temperature (100–160 °C) on structural evolution and overall film quality, examining its role during both the Bi pre-layer growth and bromination stages.
Given that Bi atomic chains form the backbone of Bi4Br4 nanoribbons, their grain size and density are closely correlated with the final size and distribution of the nanoribbons. To systematically investigate this relationship, we varied the Bi-stage growth temperature [100, 130, and 160 °C, Figs. 3(a)–(c)] while keeping the bromination temperature at a minimum to avoid interference.
In Fig. 3(a), the nanoribbons are densely distributed, with average dimensions of ~30 nm in length and ~15 nm in width. However, the low growth temperatures in both steps cause poor crystallization with amorphous regions (red circles), leading to rough surfaces and poor edge sharpness. At 130 °C [Fig. 3(b)], the higher Bi-stage growth temperature improves Bi layer quality and bromination, driving grain coalescence and enhancing b-axis growth, increasing the length to 80–100 nm (max ~140 nm) without a notable change in width, albeit with a sparser spatial distribution, and the nanoribbons with regular shapes and sharp edges reflect improved crystallinity. Large-area 3D AFM topographic images are presented in Figs. S7(a)–(c). At 160 °C [Fig. 3(c)], a further temperature increase reduces Bi density even more, severely hindering grain coalescence and growth. Consequently, short and sparse nanoribbons are observed, with the average length decreasing to ~40 nm. Since Bi deposition reduces while Br supply remains constant, the shortage of Bi reactant leads to excess Br, which forms a continuous amorphous coverage layer that is more visible on exposed substrate areas (red circles), compromising surface cleanliness. These results reveal a clear non-monotonic change in length: as the Bi-stage temperature increases, the nanoribbons first grow longer and then shorter, while their spatial density becomes progressively sparser with rising temperature.
This growth trend follows surface reaction kinetics. At low temperatures, low desorption increases Bi deposition and nucleation density. However, limited adatom mobility restricts surface diffusion, thereby hindering grain coalescence, ultimately producing small, dense grains. As the temperature rises, enhanced mobility promotes grain coalescence and longitudinal growth, significantly increasing nanoribbon length. Concurrently, higher desorption reduces the actual deposition amount, lowers nucleation density, and, together with grain coalescence, leads to a sparser distribution. At even higher temperatures, severe Bi desorption results in sparse nucleation and grain distribution that fails to coalesce, ultimately yielding short, sparsely distributed nanoribbons. Thus, the Bi-stage growth temperature significantly influences the final morphology of Bi4Br4 nanoribbons by regulating key processes like nucleation, desorption, migration, and grain coalescence.
After systematically investigating the effect of Bi temperature, we explored how bromination temperature affects the final morphology and surface quality. Appropriately increasing the bromination-stage temperature [130 °C, Figs. 3(d)–(f)] leads to more complete surface reactions and better crystallinity, reducing amorphous regions and yielding large-area nanoribbons with regular shapes and sharp edges. More importantly, higher temperatures promote desorption of unreacted Br and improve overall surface cleanliness. Figures S7(d)–(f) present the 3D topographic images of Fig. 3(e). Additional AFM images at smaller scan scales for finer surface and edge details under different growth temperatures are shown in Fig. S8.
However, excessive temperatures [160 °C, Figs. 3(g)–(i)] make Br readily desorb, hindering Bi4Br4 nanoribbon synthesis. Even when growth occurs, thermal disturbance can break their structure, causing fractures that shorten the nanoribbons and disrupt epitaxial alignment. Moreover, thermal decomposition breaks Bi–Br bonds, causing Br to volatilize and leading to a transition back to triangular Bi grains (yellow triangles). These observations reveal the significant thermal instability of the as-grown Bi4Br4 nanoribbons, as elevated temperature severely compromises the structural integrity of these 1D nanostructures. This conclusion is consistent with the annealing study in Fig. 4, together with the HRTEM findings (Figs. S2, S3) and the STM annealing results (Figs. S6).
From Fig. 3, the substrate temperature in both growth steps collectively determines the morphology (including crystal size, density and crystallinity), phase purity, and crystal orientation of the Bi4Br4 nanoribbons. Our experiments indicate that optimal film quality is achieved at well-matched intermediate temperatures: a Bi layer grown at 130 °C followed by bromination at 100–130 °C. This combination provides good Bi layer quality and suitable bromination temperatures, which benefit reaction completeness and crystallinity. Lower or higher temperatures typically cause poor crystallization, incomplete reactions or structural degradation, ultimately impairing nanoribbon structural integrity and overall film quality.
Following a systematic study of substrate temperature effects, we investigated the influence of the source deposition rate ratio on the epitaxial growth of Bi4Br4 on WTe2. Here, the Bi deposition rate and substrate temperature were held constant, while the BiBr3 deposition rate was varied to assess its impact on film quality. Our results establish the deposition rate ratio between Bi and BiBr3 as another key parameter that critically controls the final film morphology and phase purity.
At an excessively low BiBr3 deposition rate [0.6 ML min−1, Fig. 4(a)], the limited bromine supply in the reaction environment and insufficient kinetic energy restrict effective interaction with the pre-deposited Bi layer, resulting in incomplete bromination and leaving triangular Bi grains on the surface. XPS semi-quantitative analysis (Table S2) shows the Bi:Br atomic ratio of 30.92:6.13 (5:1), also indicating a large Bi excess. Within the optimal range of 1–1.2 ML min−1 [Fig. 4(b)], the supply ratio of the two sources is balanced, promoting the formation of uniform, continuous Bi4Br4 nanoribbons with well-defined 1D morphology. Further increasing the BiBr3 rate [1.5 ML min−1, Fig. 4(c)] leads to local accumulation of unreacted bromine source, forming rough amorphous regions that partially cover the nanoribbons and compromise their morphological integrity and phase purity. At 1.7 ML min−1 [Fig. 4(d)], the surface is fully covered by a continuous amorphous overlayer, wherein aligned regular features along specific crystallographic directions remain discernible, accompanied by severe degradation in both edge sharpness and overall crystallinity.
Besides the source ratio, the annealing process is equally critical in determining the final film quality. In our two-step growth procedure, the annealing temperature is generally maintained at the growth temperature, with the duration scaled proportionally to the deposition time and limited to 30 minutes. However, excessively prolonged annealing is found to trigger significant thermal decomposition. As illustrated in Fig. 4(e), extending the annealing to 45 minutes leads to the decomposition of numerous Bi4Br4 nanoribbons, characterized by Bi–Br bond cleavage and Br volatilization, resulting in the formation of triangular metallic Bi grains on the substrate. With further extension to 1 hour [Fig. 4(f)], the grain density decreases markedly, and the surface becomes nearly fully covered by triangular or spherical Bi particles, with no well-oriented Bi4Br4 nanostructures remaining. These observations clearly demonstrate the thermal instability of Bi4Br4 at elevated temperatures and highlight the importance of strictly controlled annealing duration for preserving its structural integrity.
Based on the highly controllable two-step epitaxial growth method developed in this work, we have successfully synthesized thickness-tunable Bi
4Br
4 nanostructures on WTe
2 substrates. Theoretical calculations based on the crystal structure give an out-of-plane monolayer thickness of 10.03 × sin107.4° ≈ 0.96 nm along the interlayer stacking direction, consistent with the reported value of ~0.98 nm for monolayer Bi
4Br
4 [
25], thus providing a reliable basis for thickness calibration. Through systematic optimization of growth temperature, source rate ratio, and annealing duration, we obtained large-area, uniform Bi
4Br
4 thin films [Figs. 5(a)–(d)]. By fine-tuning the deposition conditions, we fabricated nanostructures with two distinct thickness ranges: 2–4.5 nm (~2–5 layers) and 4.5–9 nm (~5–9 layers). We observed a positive correlation between increased nanoribbon thickness and both average lateral width (from ~15 nm to ~30 nm) and packing density. This yields more compact arrays, which are advantageous for integrated device fabrication. Furthermore, within localized regions of the same sample, Bi
4Br
4 nanoribbons with different layer numbers were found to coexist, offering a unique platform for systematic investigation of thickness-dependent topological properties within a single architecture.
Previous studies have reported the significant layer-number dependence of Bi
4Br
4’s electronic structure [
13,
27−
29]. In the monolayer limit, it behaves as a 2D topological insulator with a nontrivial Z
2 invariant, hosting topologically protected 1D helical edge states. As thickness increases to the bulk limit, Bi
4Br
4 transitions into a higher-order topological insulator, where symmetry-protected bulk topology gives rise to hinge states. While thinner nanoribbons may contain monolayer regions, thus exhibiting the characteristics of a 2D topological insulator, thicker ones approach bulk electronic behavior. Our thickness-tunable nanoribbons therefore enable direct observation of the cross-dimensional evolution from 2D topological insulators to quasi-1D higher-order topological insulators within a single experiment. The layer-island growth mode of Bi
4Br
4 on WTe
2 (001) further leads to the natural formation of atomically sharp step edges on the (100) surface. These steps serve as ideal 1D boundaries for exploring topologically protected dissipationless edge transport. In multilayer regions, topological boundary states from different layers exhibit weak interlayer coupling, which becomes fully decoupled when the width exceeds ~5 nm [
27], making them particularly suitable for constructing multi-channel, low-dissipation topological devices.
4 Conclusion
In summary, this work demonstrates that a two-step MBE approach using anisotropic WTe2 substrates enables the controlled, oriented growth of Bi4Br4 nanoribbons — a 1D topological insulator. Through systematic optimization of key parameters including growth temperature, source rate ratio, and annealing conditions, we achieved precise control over nanoribbon morphology, phase purity, and thickness, thereby establishing the optimal window for high-quality epitaxial growth. The resulting nanoribbons exhibit macroscopic alignment, tunable thickness from 2 to 9 nm, and well-defined morphologies with sharp edges, overcoming the persistent challenges of random orientation and poor dimensional control in conventional methods. Furthermore, this system provides an ideal platform for investigating layer-dependent topological phase transitions within a single sample. The naturally formed atomically sharp step edges and weakly coupled multilayer configurations offer unique opportunities to explore dissipationless edge transport and multi-channel topological boundary states. Our study not only presents a general strategy for the oriented integration of 1D topological materials, but also paves the way for future research on topological transport, anisotropic optoelectronic responses, and quantum device engineering, thereby advancing topological insulators from fundamental research toward functional device applications.