Synthesis of Homogeneously {100}-Textured 3-Inch Free-Standing Diamond Wafer
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Materials |
| Published | 2026-06-04 |
| Authors | J X Zhang, Stephan Handschuh‐Wang, Zhixue Xing, Tao Wang |
| DOI | 10.3390/ma19112398 |
A two-step MPCVD process produced a 3-inch free-standing diamond wafer with 1125 W/(mK) in-plane thermal conductivity. The growth rate was 4.0 to 5.5 µm/h across a ca. 175 cm² {100}-textured area. The wafer was 0.80 to 1.09 mm thick, and its conductivity exceeded 854 W/(mK) for an untextured comparison wafer.
Growth process and material conditions
Section titled “Growth process and material conditions”The authors used a commercially available, single-side polished p-type (100) silicon wafer with a 76.2 mm diameter and 3 mm thickness. Ultrasonic cleaning used acetone at ≥95.5% for 15 min, deionized water at 18.2 MΩ·cm for 5 min, and ethanol at ≥99.7% for 15 min. Nitrogen gas dried the wafer.
Ultrasonic seeding lasted 15 min in water containing 5 nm nanodiamond particles at 10 mg/mL. This electrostatic self-assembly procedure produced a seeding density of ≥10¹¹ particles/cm². The 10 kW MPCVD reactor operated at 2.45 GHz and used CH₄, H₂, and N₂ with 99.9999% purity.
The first 12 h used 400 sccm H₂, 20 sccm CH₄, and 1 sccm N₂. The nitrogen flow corresponded to 0.5 vol% in hydrogen. Growth then continued for 186 h with H₂ and CH₄ after nitrogen was stopped. The substrate temperature was 900 to 920°C. Applied power ranged from 8.5 to 9.2 kW, and pressure ranged from 11.8 to 12.0 kPa. Power and pressure were adjusted slightly to hold the sample temperature.
A 3:1 mixture of 48 wt% hydrofluoric acid and 65 wt% nitric acid detached the film from silicon and produced the free-standing wafer.
Measurement method and thickness uniformity
Section titled “Measurement method and thickness uniformity”Thickness was measured with a caliper, and growth rate was calculated by dividing final thickness by deposition time. Raman spectra used 532 nm excitation, an 1800 grating, and three averaged measurements. The authors sampled five locations along a line from one edge to the other. Optical microscopy used a 10× air objective at corresponding sites.
XRD used Cu Kα radiation over 20° ≤ 2θ ≤ 120° at three locations. A separate pole-density measurement used the same source and a χ-φ-Z sample stage. In-plane thermal diffusivity was measured by laser flash with a 40 to 50 ms pulse, 250 V laser voltage, and 17.5 mm detection area. The authors converted diffusivity to conductivity using cp = 0.54 J/(g·K) and density = 3500 kg·m⁻³.
The finished textured wafer measured 0.80 mm near the rim and 1.09 mm near the center. The center was ca. 25% thicker. The methods and Table 1 report 198 h total deposition and a growth rate of 4.0 to 5.5 µm/h. The abstract instead reports 196 h, so the total duration needs clarification.
As-grown thickness uniformity was ca. 13% for the textured sample and ca. 11% for the untextured sample. Lapping could reduce this variation further.
Texture and material quality
Section titled “Texture and material quality”Optical microscopy showed a {100}-textured surface and a crystal size of 210 ± 60 µm. The Raman diamond zone-center peak remained near 1332.0 cm⁻¹ across the plate. Its full width at half maximum increased from ca. 5 cm⁻¹ at the center to ca. 6.5 cm⁻¹ at 3 cm from the center. The Raman-derived diamond quality was 97% to 99% over the full wafer. The calculation remains approximate because its correction factor depends on excitation wavelength.
XRD patterns at all three measured locations were dominated by the diamond (400) reflection. The I400/I111 ratio ranged from 5 to 18, with an average of 11.8 ± 6.5. Pole-density data showed {100} tilt angles of 15° or less, with most facets below 10°. The crystals had no common azimuthal orientation, so the texture was uniform in plane alignment but not epitaxial. Raman intensity from non-diamond phases near 1450 cm⁻¹ rose slightly toward the center.
A continuous-nitrogen reference received 1 sccm N₂ for 180 h under otherwise similar conditions. Its thickness was 0.77 to 1.06 mm and its growth rate was 4.28 to 5.88 µm/h. Its center, about 2 cm in diameter, was dominated by (111) pyramids, while the outer 2 cm rim showed {100} texture. Rim grains measured 150 ± 50 µm, and Raman quality ranged from 88% at the center to 96% at the rim. The slightly faster reference growth did not produce a homogeneous {100} surface.
How nitrogen selected the texture
Section titled “How nitrogen selected the texture”The authors define the growth factor α as √3 times V100 divided by V111. Nitrogen raised α to about 3 during the first step. This condition favors fast {100} growth, but it can shrink {100} surfaces and leave (111) pyramids, as seen in the continuous-nitrogen reference. Stopping nitrogen sharply reduced α. The cubic grain shape led the authors to estimate a value near 1 for the second step.
Evolutionary selection explains the final surface. Faster facets overgrow slower ones during the long second step, which buries misoriented grains and leaves {111} mainly at grain boundaries. Cuboid grains present (100) planes close to parallel to the substrate.
The authors assign the 210 ± 60 µm grain size to preferential <111> growth after nitrogen was removed. They also propose that α ≤ 1 and β ≤ 1 create a broad process window for homogeneous texture. Here β compares displacement speeds of the (100) and (110) facets. In-plane misorientation prevented full crystal coalescence, leaving a polycrystalline plate with random azimuthal orientation.
Comparison with prior growth
Section titled “Comparison with prior growth”The measured 4.0 to 5.5 µm/h rate was at least four times the typical rate for MPCVD or HFCVD {100}-textured coatings. The paper places those typical rates near or below 1 µm/h. Flöter et al. reported a homogeneous textured area of about 20 cm², with crystals decreasing from 20 µm at the center to 1 to 3 µm at the rim. The present process covered ca. 175 cm² while maintaining a 210 ± 60 µm grain size across the plate.
Weng et al. previously obtained about 5 µm/h on a 3-inch substrate, but did not report texture homogeneity. The present work reports wafer-scale {100} texture uniformity and larger crystals at a similar growth rate.
Thermal performance and engineering use
Section titled “Thermal performance and engineering use”After the nucleation side was lapped and polished, the laser flash measurements gave the following in-plane results.
| In-plane measurement | {100}-textured wafer | Untextured wafer |
|---|---|---|
| Thermal diffusivity | 595 mm²/s | 452 mm²/s |
| Thermal conductivity | 1125 W/(mK) | 854 W/(mK) |
Lapping and polishing removed a few micrometers from the nucleation side and also removed the nitrogen-doped region. The authors attribute the higher conductivity to larger grains and lower non-diamond content, which reduce phonon scattering at boundaries and impurities. The nucleation side contained smaller grains and lower-quality material before removal.
Cross-plane conductivity was not measured directly. The authors estimated it as 10% to 30% above the in-plane value. Their estimates were 1237 to 1462 W/(mK) for the textured wafer and 939 to 1110 W/(mK) for the untextured wafer.
The measured in-plane conductivity remained below single-crystal diamond but agreed with reported polycrystalline diamond. For heat-spreader material selection, the data support a large-area textured plate with high in-plane conduction. Radial thickness variation and random azimuthal orientation still require consideration. The authors suggest CPU, GPU, HEMT, and high-power LED heat spreaders. Device-level performance was not reported.
Materials and support
Section titled “Materials and support”The paper calls for a free-standing {100}-textured polycrystalline diamond wafer for heat spreaders, with a 76.2 mm diameter and a 0.80 to 1.09 mm thickness. 6CCVD’s diamond materials catalog includes MPCVD-grown polycrystalline plates and wafers up to 125 mm, matching the paper’s 76.2 mm footprint. The catalog’s stated 0.1 to 500 µm thickness range is below this plate, so the engineering support link can clarify the thickness difference and the 1125 W/(mK) target for a heat-spreader design.
View Original Abstract
A two-step growth process was developed to fabricate a 3-inch homogeneously {100}-textured free-standing diamond wafer by microwave plasma enhanced chemical vapor deposition (MPCVD). The sequential growth process is based on a change in the growth parameter α, which is given by the growth rates on {100} and {111} facets, α = 3·(V100/V111). Initial growth was executed with nitrogen addition, yielding an α value close to 3 for evolutionary selection of the (100) face, followed by cessation of nitrogen addition to yield a lower α value. A homogeneously grown {100}-textured diamond over an area of ca. 175 cm2 with a thickness ≥ 0.8 mm was obtained after 196 h growth. The diamond growth rate was 4.0-5.5 µm/h, which is four times higher than the conventional growth of oriented diamond. This was substantiated by optical microscopy, Raman spectroscopy, and XRD analysis. The large crystal size of 210 ± 60 µm has been assigned to the second growth step, where growth is preferentially in the <111> direction. The homogeneous {100} texture and the large crystal size are conducive to achieving high thermal conductivity, as the in-plane thermal conductivity of the polycrystalline diamond wafer was increased from ca. 850 W/(mK) to 1125 W/(mK).
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →
Boron Doped Single Crystal Diamond Plate
View Specs →
Advanced Thermal Simulation Service Diamond Heat Spreader Performance Analysis
View Specs →References we cover
Section titled “References we cover”Original source
Section titled “Original source”References
Section titled “References”- 2024 - Applications of diamond films: A review [Crossref]
- 2024 - Enhanced Heat Dissipation of High-Power InGaN Blue Laser Diode Through Diamond Substrates [Crossref]
- 2026 - Structuring and patterning approaches for diamond—Toward the third dimension [Crossref]
- 2018 - Thermal conductivity of high purity synthetic single crystal diamonds [Crossref]
- 2014 - Relative importance of grain boundaries and size effects in thermal conductivity of nanocrystalline materials [Crossref]
- 1999 - Phonon scattering in diamond films [Crossref]
- 2014 - Large-area high-quality single crystal diamond [Crossref]
- 2003 - Thermal conductivity of epitaxially textured diamond films [Crossref]
- 2000 - High quality heteroepitaxial diamond films on silicon: Recent progresses [Crossref]
- 1993 - Textured growth of diamond on silicon via in situ carburization and bias-enhanced nucleation [Crossref]