Borophene-hBN Lateral Heterostructure

• Borophene-hBN lateral heterostructure refers to systems where borophene and hBN form adjacent domains, either via lateral stitching on Ir(111) or via step-edge epitaxy.
• Experimental insights show borophene exhibits lower friction and higher rigidity compared to hBN, as revealed by STM, nc-AFM, and tribological measurements.
• Computational studies highlight that chemically active B-terminated h-BN step edges reduce nucleation barriers, favoring one-dimensional borophene growth over two-dimensional nucleation.

Searching arXiv for the provided borophene–hBN heterostructure papers to ground the article. Borophene–hBN lateral heterostructure denotes two closely related but materially distinct configurations in the current literature. One is a laterally stitched borophene–hBN monolayer heterostructure on Ir(111) used as a platform for direct, same-tip comparison of structure, dissipation, electronic contrast, and friction between adjacent $\mathcal{X}_6$-borophene and hBN domains (Hinaut et al., 10 Jul 2025). The other is a step-edge-assisted borophene growth geometry on vicinal h-BN, in which borophene nucleates at a chemically active h-BN step edge and then extends over weakly interacting terraces (Ruan et al., 2021). The latter is described as “lateral epitaxy” in the source paper, but the computed atomic picture is not a demonstrated broad, equilibrium in-plane borophene|h-BN stitched boundary. The distinction is central: one study establishes a realized lateral comparison geometry on a metal support, whereas the other establishes an atomistic route by which h-BN step edges can seed borophene growth on an insulating support.

1. Conceptual scope and interfacial definitions

In the h-BN growth study, the active interface is a growth front anchored to an h-BN step edge, not a completed planar borophene–hBN boundary extending across two coplanar crystalline monolayers (Ruan et al., 2021). The h-BN basal plane is a wide bandgap inert insulator whose terraces are weakly interacting, while the step edges expose undercoordinated atoms that provide chemically active pinning sites. The resulting geometry is therefore a borophene sheet weakly adhering on the terrace but chemically pinned at one edge to an h-BN step. This is distinct from a canonical lateral heterostructure in which both materials occupy the same plane and are continuously bonded across an atomically narrow boundary.

By contrast, the Ir(111) study explicitly uses a borophene–hBN lateral heterostructure as an experimental platform (Hinaut et al., 10 Jul 2025). The surface contains adjacent borophene and hBN domains, total coverage close to a monolayer, and domain sizes of more than $100$ nm. The interface is lateral and sufficiently sharp to be crossed in STM, nc-AFM, KPFM, and contact-friction measurements. However, that paper also provides limited direct atomistic information about the precise atomic bonding configuration at the borophene/hBN boundary itself. Accordingly, the term “borophene–hBN lateral heterostructure” spans both an experimentally realized lateral comparison geometry on Ir(111) and a theoretically proposed edge-seeded borophene-on-h-BN architecture whose heterointerface is fundamentally step-controlled rather than fully planar.

2. Step-edge epitaxy on h-BN: atomic structure and interface chemistry

The h-BN substrate considered for borophene growth is an AA′-stacked bilayer with a step on the (0001) vicinal surface (Ruan et al., 2021). The terraces are fully saturated and chemically inert, but the step edges expose dangling-bond-like active terminations. The calculations compare B-terminated and N-terminated zigzag h-BN step edges and identify a strong asymmetry: the B-terminated zigzag edge provides good epitaxial lattice match and the lowest interfacial energy for $v1/9$ borophene, whereas growth near the N-terminated edge follows a drastically different sequence with significant lattice mismatch and rising energy. The favorable route is therefore tied specifically to a pristine B-terminated zigzag step edge, and ultrahigh vacuum or annealing may be required to avoid passivation of that edge.

The phase selected on h-BN is the $v1/9$ phase, also referred to as $\alpha$-borophene. The source attributes this to the weakly perturbing nature of the h-BN terrace: borophene adhesion on flat h-BN is nearly phase-independent and is described as being on par with the van der Waals attraction in bilayer graphene or h-BN sheets, while much weaker than borophene–Ag interaction. The Bader charge transfer from h-BN to $v1/9$ borophene is only $0.0002\,e$ per B atom, compared with about $0.03\,e$ per B atom from Ag(111) to $v1/6$ borophene, and the borophene Fermi level lies within the h-BN band gap. This means that the terrace does not electronically stabilize borophene in the way a metal can; instead, it preserves near-vacuum borophene energetics. The paper further indicates best-match periods of $2.51\ \text{\AA}$ and $100$0 for the favorable h-BN/$100$1 epitaxial relation.

The boundary character is mixed. On the terrace, the interaction is weak and vdW-like. At the step edge, by contrast, the paper repeatedly invokes strong covalent affinity between boron and the substrate step. This suggests that the relevant heterointerface chemistry is highly localized: the borophene strip is chemically stitched only at the nucleating edge, while the remainder of the sheet is supported noncovalently on h-BN.

3. Nucleation thermodynamics and growth kinetics on h-BN

The central result of the h-BN study is that clean borophene growth on inert terraces is inhibited by a large 2D nucleation barrier, whereas step edges reduce the dimensionality of nucleation to a 1D row-by-row growth mode (Ruan et al., 2021). The free-energy descriptor is written as

$100$2

with the reference choice

$100$3

The supersaturation is expressed as $100$4, i.e. the excess energy of the feedstock relative to the product borophene phase.

On flat h-BN, terrace-grown clusters evolve through compact truncated polygons that are largely triangular-lattice-rich, with occasional pentagons that later heal; the first borophene-like hollow hexagon appears only at $100$5. Along this 2D pathway, the largest Gibbs free energy explicitly reached is $100$6 at $100$7, and even at $100$8 the terrace-cluster energies exceed those on Ag(111) by 4–5 eV. Near equilibrium, the terrace nucleation barrier is effectively $100$9.

At a B-terminated h-BN step, the growth pathway changes qualitatively. Boron first forms a linear aggregate along the step; the aggregate initially covers the step with a double row and then a triple row of B atoms, illustrated by the $v1/9$0 and $v1/9$1 structures, after which the characteristic $v1/9$2 hollow-hexagon motif emerges. Growth then proceeds by repeated extension of segments such as $v1/9$3–$v1/9$4; more generally, for $v1/9$5 a segment spans $v1/9$6 boron atoms on an h-BN step-base of $v1/9$7 lattice units. In the chosen chemical-potential reference, the energy of the 1D cluster follows a near-horizontal negative line equal to the interface length times the double-edge energy of $v1/9$8 minus its binding to the h-BN step.

The kinetic precursor chosen in the nanoreactor calculations is $v1/9$9. On h-BN, a dimer is $v1/9$0 lower in energy than two monomers, the $v1/9$1 diffusion barrier is $v1/9$2 within one h-BN hexagon and $v1/9$3 across neighboring hexagons, while the monomer diffusion barrier is $v1/9$4. Substitutional trapping on the terrace is strongly disfavored for h-BN: the swap energies are $v1/9$5 for a B atom and $v1/9$6 for a dimer, whereas for graphene they are $v1/9$7 and $v1/9$8. This is why the source treats h-BN, rather than graphene, as a substrate that funnels mobile boron species toward active step sites.

The quantitative barrier contrast is decisive. The direct growth sequence gives row-initiation barriers of about $v1/9$9 for configurations such as $\alpha$0, $\alpha$1, and $\alpha$2; the effective barrier for sequential row initiation can be as low as $\alpha$3, or vanish under stronger nonequilibrium conditions. The abstract summarizes the step effect as lowering the barrier by an order of magnitude, to $\alpha$4 or less. In the supersaturation analysis, 1D growth becomes barrierless for

$\alpha$5

while the 2D terrace route still has a barrier of about $\alpha$6 under those conditions; only for

$\alpha$7

does the terrace route also become barrierless, at which point 3D boron clusters become competitive. The study therefore defines an optimal window at moderate supersaturation where 1D step-edge growth is facile but 2D terrace and 3D cluster nucleation remain suppressed. If step sites occupy about $\alpha$8 of the surface, a barrier difference of $\alpha$9 or more is estimated to ensure at least $v1/9$0 selectivity for 1D planar borophene nucleation at $v1/9$1. The corresponding precursor coverage is $v1/9$2, equivalent to a $v1/9$3 concentration of about $v1/9$4, and the study states that this is sufficient to support growth rates up to $v1/9$5; visible growth further requires a diffusion flux of order $v1/9$6 sufficient to support growth around $v1/9$7 or more.

4. Laterally stitched borophene–hBN on Ir(111): synthesis and morphology

The experimentally realized heterostructure is grown on a clean Ir(111) single crystal under UHV (Hinaut et al., 10 Jul 2025). Ir(111) preparation consists of repeated cycles of Ar$v1/9$8 sputtering at $v1/9$9 and annealing at $0.0002\,e$0. For the borophene–hBN lateral heterostructure, a diborane–borazine mixture is dosed while the substrate is kept at $0.0002\,e$1, the mixture is controlled by mass spectrometry, and annealing is maintained for $0.0002\,e$2 min after dosing. To favor pure borophene, the same precursor mixture is dosed at $0.0002\,e$3. The heterostructure is therefore formed by co-growth from a mixed precursor feed on hot Ir(111), rather than by a sequential edge-stitching protocol.

The resulting surface contains borophene islands and hBN islands, with total coverage close to a monolayer and domain sizes of more than $0.0002\,e$4. The borophene is specifically $0.0002\,e$5-borophene on Ir(111), displaying the characteristic stripe or row pattern. On larger scales, the rows reveal two of the three possible rotational domains, with $0.0002\,e$6 between row orientations. The borophene domains do not grow across Ir(111) step edges; crossing a monoatomic Ir step changes the borophene row orientation. The hBN domains form islands with the expected hexagonal moiré superstructure and can occupy different Ir terraces, including terraces adjacent to borophene.

The measured interface geometry is clear at the mesoscopic and topographic level. STM on the same terrace shows an apparent height difference of $0.0002\,e$7 at $0.0002\,e$8 between borophene and hBN. In nc-AFM topography, the specific interface shown is reported to be flat, with no height difference measured, while the Ir(111) monoatomic step under hBN appears as $0.0002\,e$9. High-resolution STM resolves the striped row pattern and a characteristic wavy appearance of $0.03\,e$0-borophene, with a measured unit cell of $0.03\,e$1 and internal angle $0.03\,e$2; high-resolution nc-AFM resolves the row structure but not the STM-visible waviness, and gives a row width of $0.03\,e$3. The study also notes that the waviness is mainly due to electronic interaction at the borophene/Ir(111) interface, whereas nc-AFM is more sensitive to topography.

Despite this detailed morphological dataset, the paper does not present a dedicated structural model of the lateral borophene/hBN boundary. It does not provide atom-resolved boundary chemistry, registry, or a DFT-relaxed borophene–hBN edge interface. Accordingly, the experimental heterostructure is structurally established as an adjacent-domain system on Ir(111), but the atomic reconstruction of the stitched boundary remains unresolved in that study.

5. Electronic contrast, dissipation, and frictional response

The lateral heterostructure on Ir(111) enables direct same-tip measurements of local electronic and tribological contrast (Hinaut et al., 10 Jul 2025). KPFM/CPD mapping clearly distinguishes borophene and hBN: borophene has higher CPD than hBN, with a work-function difference of $0.03\,e$4. Using the Ir(111) work function of $0.03\,e$5 as reference, the study gives $0.03\,e$6 and $0.03\,e$7. The paper notes that this work-function contrast may influence electronic dissipation channels, although it does not establish that effect as dominant.

nc-AFM dissipation imaging shows a pronounced mechanical contrast. hBN displays much higher dissipation than borophene, the hBN moiré is visible in dissipation, and hBN islands on different terraces have identical dissipation. Borophene dissipation is lower overall; on borophene, the lowest dissipation occurs on the rows, while higher dissipation between rows may be related to defects or adsorbates. The reported dissipation ratios are $0.03\,e$8 and $0.03\,e$9. The interpretation given is that borophene is flatter, more rigid, and more strongly adhered to Ir, whereas hBN’s moiré structure is more deformable under the oscillating tip.

The frictional contrast is the central empirical result. For $v1/6$0-borophene, the calculated coefficient of friction at low loads is $v1/6$1, and superlubric sliding persists up to $v1/6$2 load. Above this threshold, friction becomes nonlinear and superlubricity disappears; after reducing the load below threshold, superlubricity is reversibly restored. For hBN measured with the same tip, the coefficient of friction is $v1/6$3. The source therefore identifies hBN as low-friction, but clearly less lubricious than borophene in this system. A particularly strong demonstration scans across an hBN island embedded between two $v1/6$4-borophene domains of different orientation: on the borophene sides, forward and backward traces are similar and show no hysteresis; over the hBN middle region, the traces show clear hysteresis and increased friction. Atomic stick-slip is resolved on borophene at low load and near the transition load, but in the superlubric regime this atomic corrugation is not accompanied by forward/backward hysteresis. The representative scan speed reported for borophene lateral-force traces is $v1/6$5. The conclusion further states that hBN exhibits higher friction and remains wear-free.

6. Modeling frameworks, unresolved questions, and implications for true borophene–hBN lateral junctions

The two studies use different theoretical frameworks and leave different open questions. The h-BN work is purely computational, based on DFT-D2/PBE calculations using PAW in VASP with a $v1/6$6 plane-wave cutoff and a $v1/6$7 vacuum slab (Ruan et al., 2021). Its strength is mechanistic: it identifies why an inert insulating terrace suppresses borophene nucleation, why a B-terminated zigzag step edge changes the nucleation dimensionality, and why the resulting borophene should remain electronically decoupled and readily transferable. Its limitation is equally explicit: it does not provide a model of an extended coplanar borophene domain covalently fused to an h-BN monolayer domain across a single atomically sharp interface, nor interfacial electronic states, band offsets, or transport across such a junction.

The Ir(111) study couples experiment to a modified Prandtl–Tomlinson model and to separate DFT calculations for the two materials (Hinaut et al., 10 Jul 2025). The PT formulation extends the standard

$v1/6$8

description by introducing a deformable surface coordinate, with total potential

$v1/6$9

and tip–surface interaction

$2.51\ \text{\AA}$0

The model uses the experimental lattice parameters $2.51\ \text{\AA}$1, $2.51\ \text{\AA}$2, and $2.51\ \text{\AA}$3; the equations are solved by 4th-order Runge–Kutta, average friction is obtained over 82 cycles, and damping is set to critical damping. The key fitted result is $2.51\ \text{\AA}$4.

The accompanying ab initio calculations use two separate vertical stacks, Ir/$2.51\ \text{\AA}$5-borophene/Si tip and Ir/hBN/Si tip, not a lateral boundary model. The setup employs VASP, $2.51\ \text{\AA}$6 vacuum along $2.51\ \text{\AA}$7, PBE for borophene, vdW-DF2 for hBN, a $2.51\ \text{\AA}$8 cutoff, a $2.51\ \text{\AA}$9 Monkhorst–Pack mesh, $100$00 SCF convergence, $100$01 geometry convergence, and a $100$02 PES grid. The sliding barrier is $100$03 for borophene and $100$04 for hBN, while the average vertical displacement during sliding is $100$05 for borophene and $100$06 for hBN. The paper interprets the lower borophene friction in terms of weaker tip–surface interaction corrugation, smaller sliding barrier, greater out-of-plane rigidity, and stronger adhesion to Ir(111), which suppresses puckering-mediated dissipation.

Taken together, the two studies delimit what is and is not presently established. A laterally stitched borophene–hBN monolayer heterostructure on Ir(111) has been used successfully as a direct comparison platform, but its atomic boundary reconstruction is not resolved. A step-edge-anchored borophene-on-h-BN geometry has been analyzed in detail on an insulating support, but it is not a demonstrated broad in-plane borophene|h-BN junction. This suggests several design principles for a true borophene–hBN lateral heterostructure: chemically active, unpassivated B-terminated zigzag edges; inert terraces with low substitutional trapping; moderate supersaturation that favors 1D initiation while suppressing terrace 2D nucleation and 3D clustering; and, plausibly, a geometry in which borophene grows from the sidewall or monolayer edge of h-BN, rather than merely along a step on top of a multilayer support. Within the literature considered here, those features are implied rather than demonstrated.