Carbon Buffer Layer-Assisted Confinement Epitaxy

• Carbon buffer layer-assisted confinement epitaxy is a materials engineering approach that uses atomically thin carbon films to guide epitaxial growth by modulating interface chemistry, strain, and atomic registry.
• This method suppresses undesired step bunching and interdiffusion, enabling the synthesis of atomically sharp heterostructures for 2D materials and ultrathin semiconductors.
• Optimized buffer characteristics such as controlled thickness and low roughness allow precise strain modulation and nucleation control essential for scalable high-quality epitaxial films.

Carbon buffer layer-assisted confinement epitaxy is a materials engineering strategy in which atomically thin carbon films—specifically, reconstructed graphene-like “buffer layers” or tailored amorphous carbon interlayers—mediate the growth of epitaxial thin films by modulating interface chemistry, strain, and atomic registry. This approach controls the interface energetics, suppresses undesired step bunching or interdiffusion, and enables synthesis of high-quality, atomically sharp heterostructures, especially for two-dimensional (2D) materials and ultrathin metals or semiconductors on polar or lattice-mismatched substrates. The paradigm applies most notably to graphene growth on SiC(0001), but extends to advanced confinement heteroepitaxy of 2D metals (e.g. Ga, In, Sn) and remote epitaxy of III–V or III–N compounds on functionalized amorphous carbon layers.

1. Structural and Chemical Nature of Carbon Buffer Layers

On SiC(0001), a canonical carbon buffer layer (“zero-layer graphene,” ZLG) forms via surface graphenization: it is a graphene-like honeycomb monolayer whose lattice reconstructs into a $6(\sqrt{3}\times\sqrt{3})R30^\circ$ supercell, rotated by 30° with respect to the SiC lattice. Approximately one third of the C atoms in ZLG form sp³ bonds with topmost Si atoms, producing a long-range (6×6) corrugation with Moiré periodicity (Schumann et al., 2014). Scanning tunneling microscopy (STM) shows a peak-to-peak corrugation $\Delta h_{\mathrm{BL(pp)}} \approx 1.0$ Å and root-mean-square (RMS) roughness of $0.18\pm 0.02$ Å (Goler et al., 2011). The reconstruction stabilizes the interface, imposes atomic registry, and introduces localized pseudo-gap features due to partial $\pi$-network disruption (Goler et al., 2011).

Alternately, plasma-enhanced chemical vapor deposition (PECVD) enables growth of ultrathin amorphous carbon (a-C) templates with controlled thickness (0.1–2 nm), RMS roughness $\leq$ 0.3 nm, and predominantly sp² bonding ($f_{\mathrm{sp}^2} \approx 0.90$ for $t_{a\text{-}C}=10$ s growth). Such a-C layers act as “universal” interlayers for remote/confinement epitaxy on III–V and III–N substrates (Henksmeier et al., 2024).

2. Mechanisms of Confinement and Interface Control

The confinement effect arises from the combination of chemical bonding, mechanical rigidity, and electronic structure of the carbon buffer. On SiC, sp³ C–Si anchors in the buffer rigidly pin the carbon lattice and align subsequent epitaxial overlayers to the substrate, while confining mobile adatoms and suppressing desorption [(Schumann et al., 2014); (Goler et al., 2011)]. This “template” effect not only fixes orientational registry (eliminating rotational domains) but also laterally confines mass transport, leading to stable step terraces and marked suppression of step bunching during high-temperature sublimation growth (Kruskopf et al., 2017). The reconstructed buffer locally regulates chemical potential gradients ($\Delta\mu$) and step line tension ($\gamma$), resulting in refined step kinetics described by $v = (\Omega D / kT) \cdot (\Delta\mu / \lambda)$ for step velocity (Kruskopf et al., 2017).

In amorphous carbon–assisted cases, the in-plane sp² network partially screens substrate potential in the growth direction but transmits lateral registry. The substrate–film interaction decays exponentially through the carbon interlayer as $\Delta h_{\mathrm{BL(pp)}} \approx 1.0$0 with $\Delta h_{\mathrm{BL(pp)}} \approx 1.0$1 (Henksmeier et al., 2024). By adjusting $\Delta h_{\mathrm{BL(pp)}} \approx 1.0$2, one tunes between strong and weak confinement regimes, with direct implications for nucleation barrier ($\Delta h_{\mathrm{BL(pp)}} \approx 1.0$3), dislocation density, and strain transfer ($\Delta h_{\mathrm{BL(pp)}} \approx 1.0$4 interface stretching for GaAs/a-C) (Henksmeier et al., 2024).

3. Quantitative Lattice and Strain Modulation

Grazing-incidence X-ray diffraction (GID) and Raman measurements yield precise lattice parameters for carbon buffer-confined systems. On SiC(0001):

System	Lattice constant (Å)
Uncovered buffer layer ($\Delta h_{\mathrm{BL(pp)}} \approx 1.0$5)	2.467
Interfacial BL beneath monolayer graphene	2.463
Monolayer graphene on BL ($\Delta h_{\mathrm{BL(pp)}} \approx 1.0$6)	2.456
Quasi–free-standing bilayer graphene (BLG)	2.460
Strain-free graphite ($\Delta h_{\mathrm{BL(pp)}} \approx 1.0$7)	2.461

The interfacial BL is primarily responsible for compressive strain in the overlying monolayer graphene, with $\Delta h_{\mathrm{BL(pp)}} \approx 1.0$8 given by $\Delta h_{\mathrm{BL(pp)}} \approx 1.0$9—reflecting the "confined" state induced by pinning and lattice mismatch (Schumann et al., 2014). Upon intercalation (e.g., oxygen, metal), breaking of sp³ bonds decouples the buffer, relaxing the overlayer and converting the BL into a second graphene layer ($0.18\pm 0.02$0 Å, $0.18\pm 0.02$10.04% compressive residual) (Schumann et al., 2014). This mechanics underlies the strain-engineering capabilities enabled by buffer layer control.

In amorphous carbon–mediated epitaxy, interface strain is likewise quantifiable: for GaAs/a-C heterointerfaces, geometric phase analysis of STEM images yields $0.18\pm 0.02$2–$0.18\pm 0.02$3 (Henksmeier et al., 2024).

4. Advanced Confinement Heteroepitaxy and Intercalation

The buffer layer at the graphene/SiC interface enables “confinement heteroepitaxy” (CHet) for direct synthesis of atomically thin elemental metals (Ga, In, Sn) stabilized by the buffer’s dual role as a structural cap and template (Briggs et al., 2019, Mamiyev et al., 18 Feb 2026). Intercalation proceeds by diffusion of metal atoms beneath the carbon buffer, driven by chemical potential gradients and activation barriers reduced by the buffer layer (e.g., for Sn: $0.18\pm 0.02$4 eV, $0.18\pm 0.02$5 eV) (Mamiyev et al., 18 Feb 2026). CHet products show an explicit vertical bonding gradient: covalent substrate–metal interface, metallic/covalent bonding between confined metal layers, and vdW or weak graphene–metal capping (Briggs et al., 2019).

For Sn, the process yields a (1×1)-registered 2D triangular Sn layer between SiC and quasi-free-standing graphene, with in-plane strain relaxation and stabilization of exotic phases unattainable via direct deposition (Mamiyev et al., 18 Feb 2026). Metal intercalation under buffer graphene is also integral to scalable, high-fidelity QFMLG production and heterostructure engineering.

5. Functional Electronic, Superconducting, and Strain-Engineering Outcomes

The carbon buffer layer’s action as a confinement medium has significant functional consequences:

• Graphene/SiC platforms: The ability to modulate strain, orientation, and defect density of graphene via buffer coupling is vital for tuning phonon, electronic, and quantum transport properties (Schumann et al., 2014).
• Remote epitaxy of III–V/III–N: Atomically thin a-C templates with tailored thickness yield high-quality pseudomorphic films (dislocation densities $0.18\pm 0.02$6, smooth surfaces) and facilitate easy epitaxial lift-off for membrane integration (Henksmeier et al., 2024).
• Half–van der Waals metals and quantum heterostructures: Buffer-assisted CHet enables realization of atomically thin Ga, In, Sn, with experimentally verified Fermi velocities ($0.18\pm 0.02$7 up to $0.18\pm 0.02$8 m/s), symmetry breaking, and novel superconductivity (Briggs et al., 2019, Yi et al., 6 Sep 2025).
• Emergent superconductivity: Trilayer Ga grown by carbon-buffer-layer–assisted confinement shows interfacial Ising-type superconductivity, with upper critical fields $0.18\pm 0.02$9 the Pauli limit due to strong orbital hybridization and spin–orbit coupling induced by the buffer-mediated interface (Yi et al., 6 Sep 2025).
• Strain-Engineering: Dynamical coupling between decoupled QFMLG and the underlying confined metal allows strain tuning via thermal expansion coefficient mismatch and epitaxial registry, as monitored by temperature-dependent Raman shifts (Mamiyev et al., 18 Feb 2026).

6. Fabrication Protocols, Optimization, and Scalability

Key process variables and optimization strategies are established for carbon buffer-layer–assisted confinement epitaxy:

• Buffer formation on SiC: Employ H-etch/anneal (e.g. 1175–1400 °C), optional polymer-assisted C-supply to seed uniform buffer islands, followed by high-temperature graphenization (Kruskopf et al., 2017).
• Amorphous carbon template: PECVD at 300 °C (ICP 300 W, RF 10 W) achieves monolayer a-C ($\pi$0 nm) for maximum lattice registry and low-defect density remote epitaxy (Henksmeier et al., 2024).
• Intercalation/CHet protocols: O₂/He plasma defect engineering, controlled vapor-phase metal intercalation at 600–800 °C, and proper anneal/cooldown cycles are essential for controlling interfacial order, lattice matching, and phase quality (Briggs et al., 2019).
• Thickness and registry control: For a-C, optimized $\pi$1 nm is necessary to sustain epitaxial pull-through and minimize dislocations; excessive thickness decouples the registry and increases defectivity (Henksmeier et al., 2024).
• Scalability: Carbon buffer approaches are compatible with wafer-scale SiC substrates, allowing integration into device fabrication workflows and patterning for quantum/spintronic applications (Yi et al., 6 Sep 2025).

7. Outlook, Generalizations, and Prospects

Carbon buffer layer-assisted confinement epitaxy generalizes across:

• Substrate classes: from polar/bulk SiC(0001) to III–V and III–N on silicon or nitride substrates (Henksmeier et al., 2024).
• Buffer chemistries: both crystalline ZLG (graphene buffer) and amorphous sp²-rich carbon serve as tunable mediators for confinement and registry, with the potential to extend to hBN or other 2D dielectrics (Mamiyev et al., 18 Feb 2026).
• Intercalated/epitaxial phases: from graphene and QFMLG to confined metals, superconductors, and metastable semiconductors, with unique opportunities in symmetry breaking, electronic topology, and quantum allocation (Briggs et al., 2019, Yi et al., 6 Sep 2025).

A plausible implication is that modular manipulation of carbon buffer chemistry, morphology, and thickness, in conjunction with controlled intercalation/annealing cycles, will remain foundational for the synthesis and engineering of low-dimensional materials and hybrid quantum platforms.

---

References

• (Schumann et al., 2014, Goler et al., 2011, Kruskopf et al., 2017, Briggs et al., 2019, Henksmeier et al., 2024, Mamiyev et al., 18 Feb 2026, Yi et al., 6 Sep 2025)