Study of cracks formation in HIGHLY – low boron-doped epitaxial (113) diamond bilayers
At a Glance
Section titled “At a Glance”| Metadata | Details |
|---|---|
| Publication Date | 2021-01-01 |
| Journal | NANOCOM … |
| Authors | V. Mortet, Ladislav Klimša, N. Lambert, Marina Davydova, Jaromı́r Kopeček |
| Institutions | Czech Academy of Sciences, Institute of Physics, Czech Technical University in Prague |
| Citations | 1 |
| Analysis | Full AI Review Included |
Executive Summary
Section titled “Executive Summary”- Core Problem Addressed: The formation of cracks in highly boron-doped (p+) / undoped (i) epitaxial diamond bilayers, which prevents the reliable fabrication of semi-vertical diamond Schottky diodes.
- Material System: Epitaxial diamond layers grown on (113) oriented substrates using Microwave Plasma Enhanced Chemical Vapor Deposition (MWPECVD). The (113) orientation is preferred for its superior surface morphology and high boron incorporation efficiency.
- Critical Finding (Thickness): A critical thickness (tcrit) of approximately 3.5 µm was experimentally observed for the undoped (i) diamond layer. Layers grown thicker than this threshold consistently exhibited cracking.
- Mechanism of Cracking: Crack formation is attributed to the relaxation of mechanical energy stored in the undoped layer due to a significant lattice mismatch (ca. 0.8%) between the highly boron-doped (p+) layer and the subsequent undoped layer.
- Doping Level: The underlying p+ layer was ca. 5 µm thick with a high boron concentration of ca. 1021 cm-3, yielding a low resistivity (2 mΩ.cm) suitable for ohmic contacts (< 10-6 Ω.cm2).
- Process Independence: The critical thickness was found to be independent of the methane concentration (varied from 0.3% to 1%) used during the undoped layer deposition.
Technical Specifications
Section titled “Technical Specifications”| Parameter | Value | Unit | Context |
|---|---|---|---|
| Substrate Orientation | (113) | N/A | High-pressure high-temperature (HPHT) diamond |
| Growth Method | MWPECVD (AX5010) | N/A | Microwave Plasma Enhanced Chemical Vapor Deposition |
| Highly B-Doped Layer Thickness (1st step) | ca. 5 | µm | Fixed thickness |
| Highly B-Doped Concentration ([B]) | ca. 1021 | cm-3 | Determined by Raman spectroscopy |
| B/C Ratio (1st step) | 2000 | ppm | Gas phase ratio |
| Undoped Layer Critical Thickness (tcrit) | ca. 3.5 | µm | Threshold for crack formation |
| Lattice Mismatch (p+ vs. i layer) | ca. 0.8 | % | Primary cause of stored elastic energy |
| Highly B-Doped Resistivity | 2 | mΩ.cm | Suitable for ohmic contact formation |
| Target Specific Contact Resistance | < 10-6 | Ω.cm2 | Required for low ON resistance Schottky diodes |
| Undoped Layer CH4 Concentration (2nd step) | 0.3 to 1 | % | Varied parameter; did not affect tcrit |
| Deposition Pressure | 100 | mbar | Constant for both steps |
| Microwave Power | 700 | W | Constant for both steps |
| Deposition Rate (1% CH4) | ca. 2.5 | µm/h | Estimated growth rate |
| Raman Excitation Wavelength | 488 | nm | Used for boron concentration analysis |
Key Methodologies
Section titled “Key Methodologies”The epitaxial bilayers were grown in a two-step process using a commercial AX5010 MWPECVD reactor:
-
Step 1: Highly Boron-Doped (p+) Layer Deposition
- Goal: Create a low-resistivity layer for ohmic contact.
- Recipe: 100 mbar pressure, 700 W microwave power, 700 sccm total gas flow.
- Gas Phase: 1% CH4, 2000 ppm Boron-to-Carbon (B/C) ratio (using trimethylborane precursor).
- Duration/Result: 2 hours, resulting in a ca. 5 µm thick layer with [B] ≈ 1021 cm-3.
-
Step 2: Undoped (i) Layer Deposition
- Goal: Grow the intrinsic layer required for the Schottky diode structure.
- Recipe: Boron precursor flow was immediately stopped. Pressure and power remained constant (100 mbar, 700 W).
- Gas Phase: 0% B/C, CH4 concentration varied between 0.3% and 1%.
- Duration/Result: Variable deposition time to achieve different thicknesses (t).
Characterization and Analysis:
- Thickness Determination: Estimated based on known deposition rates derived from SIMS analysis of reference multilayer samples.
- Boron Concentration Measurement: Performed using Raman spectroscopy (488 nm laser) by fitting the decoupled double Fano-function to the spectrum, specifically analyzing the width of the zone center phonon (ZCP) line.
- Crack and Defect Observation: Determined using optical microscopy (Zeiss Imager Z1m) and high-resolution Scanning Electron Microscopy (SEM, TESCAN FERA3 GM).
Commercial Applications
Section titled “Commercial Applications”This research directly supports the development and manufacturing of diamond-based semiconductor devices, leveraging diamond’s extreme material properties.
- High-Power Electronics: Fabrication of robust, high-voltage switching devices, including Schottky diodes and MOSFETs, utilizing diamond’s high breakdown field and thermal conductivity.
- High-Frequency (RF) Devices: Diamond’s high carrier mobility and saturation velocity make it ideal for high-frequency applications where thermal management is critical.
- Semi-Vertical Device Architecture: The specific p+/i bilayer structure is fundamental to creating semi-vertical diodes, which offer advantages in current handling and device footprint compared to lateral designs.
- Wide-Bandgap Semiconductor Technology: Advancing the fundamental understanding of strain engineering and defect mitigation in highly mismatched epitaxial diamond structures.
View Original Abstract
In this work, we present the study of the formation of cracks in high and low boron-doped diamond epitaxial bilayers necessary in the fabrication process of Schottky diodes.Epitaxial diamond layers were grown on (113) oriented diamond substrates by Microwave Plasma Enhanced Chemical Vapor Deposition.The effect of the thickness and the methane concentration during the growth of the undoped diamond layer on the crack formation have been studied using optical and scanning electron microscopy (SEM).We experimentally observed a critical thickness of ca.3.5 µm above which all undoped layers are cracked.The formation of these cracks is attributed to the relaxation of the elastic energy stored in the epitaxial undoped layer due to the significant lattice mismatch (ca.0.8 %) between the undoped and highly boron-doped diamond layers with a boron concentration of 10 21 cm -3 as determined by Raman spectroscopy analysis.