Optimization for growth condition of ultrathin hexagonal boron nitride on dielectric substrates via LPCVD

Hexagonal Boron Nitride (h-BN) is a highly intriguing candidate for heterostructure optoelectronic applications, such as Deep Ultraviolet photodetectors, UV sensing and communication systems and solar cells. This is primarily due to its unique properties, including a layer dependent wide energy bandgap, superior mechanical strength, high thermal conductivity, high band-edge absorption coefficient, and exceptional transparency in the UV region. The widely adopted synthesis method for h-BN thin films is Chemical Vapor Deposition (CVD) Method, which often utilizes catalytic substrates like copper (Cu) and Nickel (Ni). However, integrating the synthesized h-BN into device applications requires a subsequent transfer process to the target substrate. This transfer step introduces significant material damage, such as folding, cracking and polymer residues, which ultimately degrade the optoelectronic properties of the material and compromise device performance. To overcome this major challenge, there is a strong need to synthesize high-quality h-BN films directly onto dielectric substrates such as silicon (Si), SiO2, quartz, sapphire or AlN without the need for transfer. The primary difficulty in direct synthesis lies in achieving homogenous, high crystallinity films with controllable thickness due to the absence of a catalytic effect. In this work, we investigated the optimization of growth parameters for the direct synthesis of ultrathin h-BN films on non-catalytic quartz substrates, which are highly transparent in the UV region, using the Low-pressure Chemical Vapor Deposition (LPCVD) method. The optimal synthesis conditions were determined to be 1050oC for 60 min, achieved by the decomposition of 150 mg Ammonia Borane (AB) precursor at 80oC. This optimization is crucial for advancing large-scale, high-performance h-BN based DUV photodetector fabrication.

💡 Research Summary

The paper addresses a critical bottleneck in the integration of hexagonal boron nitride (h‑BN) into deep‑ultraviolet (DUV) optoelectronic devices: the need to transfer CVD‑grown films from catalytic metal substrates (Cu, Ni) to dielectric target substrates, a step that introduces cracks, folds, polymer residues, and other defects that degrade device performance. To eliminate the transfer step, the authors explore the direct synthesis of ultrathin h‑BN on a non‑catalytic, UV‑transparent dielectric—quartz—using low‑pressure chemical vapor deposition (LPCVD).

A systematic parametric study is performed. The key variables are growth temperature (1000–1050 °C), growth time (15–90 min), precursor amount (50–200 mg of ammonia‑borane, AB), and AB decomposition temperature (80–100 °C). The LPCVD reactor is configured with a 120‑cm quartz tube, 300 sccm of 10 % H₂/Ar carrier gas (270 sccm Ar, 30 sccm H₂), and an initial pressure of 6.5 × 10⁻¹ mbar. The substrates are cleaned ultrasonically in acetone, IPA, and DI water, then loaded 50 cm downstream of the AB source. The furnace is ramped at 10 °C min⁻¹ to the target temperature, held for the prescribed time, and finally cooled under a continuous flow of 500 sccm 10 % H₂/Ar.

X‑ray photoelectron spectroscopy (XPS) is used to monitor the chemical state of the films. The B 1s peak at ~191 eV and the N 1s peak at ~398.5 eV confirm sp²‑bonded h‑BN. As growth time increases, the intensities of both peaks rise linearly, while the Si 2p signal from the underlying quartz attenuates, indicating progressive coverage and thickness control. The narrowest full‑width‑half‑maximum (FWHM) values—2.98 eV for B 1s and 2.28 eV for N 1s—are observed for a 60‑minute growth, signifying the highest crystallinity among the tested durations.

Varying the AB mass reveals that 150 mg yields the optimal stoichiometry (B:N ≈ 1.03) and the smallest peak widths (B 1s 2.49 eV, N 1s 2.27 eV). Lower amounts (≤ 50 mg) lead to insufficient precursor flux and thinner, less uniform films; higher amounts (≥ 200 mg) risk B‑rich non‑stoichiometric regions. The data suggest that the process operates in a precursor‑limited kinetic regime, where increasing solid AB directly raises the gas‑phase concentration of active B‑N species, thereby accelerating growth without compromising film quality.

The temperature study confirms that 1050 °C, combined with a decomposition temperature of 80 °C for AB, provides enough surface activation on quartz to promote nucleation and lateral growth despite the absence of a metallic catalyst. The chosen heating rate (10 °C min⁻¹) and carrier‑gas composition ensure a stable low‑pressure environment that suppresses unwanted gas‑phase nucleation and promotes uniform film formation.

In summary, the authors identify the optimal LPCVD recipe for direct h‑BN growth on quartz: 1050 °C, 60 min, 150 mg AB decomposed at 80 °C, under 300 sccm 10 % H₂/Ar at 6.5 × 10⁻¹ mbar. This condition yields ultrathin, high‑crystallinity h‑BN with near‑ideal stoichiometry, minimal defect density, and controllable thickness—all without a transfer step. The work demonstrates that high‑quality 2D insulating layers can be fabricated directly on dielectric substrates, opening a scalable pathway for integrating h‑BN into DUV photodetectors, UV sensors, and other heterostructure devices where transparency, thermal stability, and mechanical robustness are essential. Future directions include scaling the process to wafer‑scale, investigating other dielectric substrates (SiO₂, sapphire, AlN), and integrating the films into functional DUV devices to benchmark performance gains over transferred h‑BN.