PC SiC Single Crystal Coefficient of Thermal Expansion Measurement Experiment

The coefficient of thermal expansion (CTE) of PC silicon carbide (SiC) single crystals is a critical parameter for power device design, epitaxial growth control, and wafer-level thermal stress modeling. This article presents a complete technical overview of the CTE measurement experiment for PC SiC single crystals, covering sample preparation, testing items, measurement methods based on ASTM and ISO standards, typical results, and industrial applications. Readers will gain a practical understanding of how anisotropic thermal expansion data for 4H-SiC and 6H-SiC crystals are obtained, validated, and applied.

Technical Overview

Silicon carbide single crystals, including the PC SiC grade used for power semiconductor substrates, exhibit an exceptionally low and highly anisotropic coefficient of thermal expansion (CTE). Because SiC wafers routinely undergo high-temperature processes such as epitaxy at approximately 1,600 °C, ion-implantation annealing, oxidation, and die attachment, accurate CTE data are essential for predicting wafer warpage, film stress, and device reliability. The purpose of the PC SiC single crystal thermal expansion coefficient measurement experiment is to quantify the linear thermal expansion behavior of the crystal along both the basal-plane (a-axis) and crystallographic (c-axis) directions as a function of temperature.

SiC crystallizes in a hexagonal structure for the technologically important 4H and 6H polytypes, which means the thermal expansion is direction-dependent. Typical reported values at room temperature are approximately αa ≈ 4.2 × 10⁻⁶ /K and αc ≈ 4.7 × 10⁻⁶ /K, and both coefficients increase gradually with temperature. The experiment described here determines the mean CTE, the instantaneous CTE, and the full expansion curve from room temperature up to 1,000–1,600 °C, with a measurement uncertainty typically within ±0.1 × 10⁻⁶ /K when calibrated dilatometry is used.

• Determines mean and instantaneous linear CTE of PC SiC single crystals
• Resolves anisotropic expansion along the a-axis and c-axis
• Covers the temperature range from 25 °C to 1,600 °C in inert atmosphere
• Follows recognized standards such as ASTM E228, ASTM E831, and ISO 11359-2

Testing Samples

PC 4H-SiC Single Crystal Specimens

4H-SiC is the dominant polytype for power MOSFETs and Schottky barrier diodes. Samples are prepared from as-grown boules or finished wafers by cutting rectangular bars or cylindrical rods with a diamond wire saw, followed by grinding and polishing to achieve parallel end faces. Orientation is verified by X-ray diffraction prior to testing.

• Typical dimensions: 3 mm × 3 mm × 10 mm bars or 6 mm diameter × 20 mm rods
• End-face parallelism within ±5 µm to minimize dilatometer alignment error
• Surface roughness Ra ≤ 0.8 µm after fine grinding
• Crystallographic orientation error within ±0.5°

PC 6H-SiC Single Crystal Specimens

6H-SiC specimens are used for applications in RF electronics and high-temperature sensors. Sample preparation follows the same machining flow as 4H-SiC. Because the c/a ratio of 6H-SiC differs from 4H-SiC, separate CTE data must be measured rather than extrapolated.

Orientation-Specific Samples for Anisotropy Evaluation

To fully characterize the anisotropic thermal expansion, paired specimens are machined with the measurement axis parallel to the a-axis and parallel to the c-axis respectively. For on-axis (0001) wafers, in-plane bars supply the a-axis coefficient, while c-axis rods are extracted from the boule along the growth direction.

• a-axis specimen: measurement direction along [11-20]
• c-axis specimen: measurement direction along [0001]
• Reference specimen of fused silica or sapphire run in parallel for drift correction

Testing Items

The experiment is designed to deliver a complete thermal expansion dataset for PC SiC single crystals. The following quantities are measured and reported:

• Mean linear coefficient of thermal expansion (αm) over defined intervals such as 25–200 °C, 25–500 °C, 25–1,000 °C, and 25–1,600 °C
• Instantaneous (differential) CTE as a continuous function of temperature
• Anisotropic coefficients αa and αc and their ratio as a function of temperature
• Fractional length change ΔL/L0 versus temperature curves for heating and cooling segments
• CTE hysteresis and repeatability after multiple thermal cycles
• Temperature-dependent lattice expansion parameters a(T) and c(T) when the X-ray diffraction method is applied

Testing Methods

Push-Rod Dilatometry (ASTM E228)

The primary method uses a horizontal push-rod dilatometer with a fused silica or high-purity alumina holder. The specimen length change is transmitted through a low-load push rod to a linear variable differential transformer (LVDT) with nanometer-level resolution. The fundamental calculation is expressed as α = ΔL / (L0 × ΔT), where L0 is the initial sample length at the reference temperature.

• Calibration is performed against NIST standard reference materials such as fused silica (SRM 739) or sapphire (SRM 732)
• Heating rate of 3–5 K/min under high-purity argon to suppress surface oxidation above 600 °C
• Holder expansion is automatically subtracted using a blank baseline run
• Typical temperature range: room temperature to 1,000 °C with silica system, up to 1,600 °C with alumina system

Thermomechanical Analysis (ASTM E831)

For limited quantities of wafer material, thermomechanical analysis permits CTE determination on small specimens a few millimeters in length. A quartz probe applies a static force of 0.05–0.2 N while the furnace ramps at 5 K/min. This method is well suited for comparing expansion behavior across wafer positions and doping conditions.

Laser Interferometric Dilatometry

For the highest precision, a dual-path Michelson interferometer measures the displacement of the specimen end faces with resolution approaching 10 nm. This technique is preferred when distinguishing small differences between αa and αc at room temperature, where the absolute expansion over a 100 K interval is only a few hundred nanometers.

High-Temperature X-Ray Diffraction Lattice Parameter Method

Because dilatometry measures bulk length change, the anisotropic lattice expansion is independently verified by high-temperature XRD. Lattice parameters a and c are refined at temperature steps of 100 K, and the coefficients are derived as αa = (1/a)(da/dT) and αc = (1/c)(dc/dT). Agreement between dilatometry and XRD within 2% confirms the absence of microcracking or residual stress relaxation artifacts.

Applications

Accurate CTE data for PC SiC single crystals directly support multiple stages of power electronics manufacturing and reliability engineering:

• Epitaxial growth modeling, where CTE mismatch between the SiC substrate and the epilayer determines cooling-induced bow and defect generation
• Wafer processing simulations for oxidation, annealing, and laser dicing that require temperature-dependent expansion input
• Die-attach and packaging design, where the SiC CTE of approximately 4.2–4.7 × 10⁻⁶ /K governs solder joint and sintered-silver layer stress compared with copper and aluminum nitride heat spreaders
• GaN-on-SiC RF device design, where expansion mismatch drives channel stress and cracking risk
• Quality control of crystal growth, since abnormal expansion behavior can reveal polytype inclusions, micropipes, or residual stress in the boule
• Finite element thermal-mechanical modeling for automotive and aerospace inverters operating under wide temperature swings

FAQ

What is the coefficient of thermal expansion of PC SiC single crystal?

For hexagonal 4H and 6H SiC single crystals, the mean linear CTE near room temperature is approximately 4.2 × 10⁻⁶ /K along the a-axis and 4.7 × 10⁻⁶ /K along the c-axis. The mean value over the range of 25 °C to 1,000 °C typically falls between 4.5 × 10⁻⁶ /K and 5.2 × 10⁻⁶ /K, rising gradually with temperature.

Why must a-axis and c-axis samples be measured separately?

SiC single crystals are hexagonal and therefore thermally anisotropic. A single in-plane measurement cannot represent the c-axis expansion, which is 10–15% larger. Both orientations must be characterized to build correct anisotropic material models for stress simulation.

Which standards govern the experiment?

Push-rod dilatometry is performed per ASTM E228, thermomechanical analysis per ASTM E831, and the general plastics/thermomechanical dilatometry approach per ISO 11359-2. Cross-verification by high-temperature XRD follows standard crystallographic refinement practice.

What atmosphere is used, and why?

Tests above 600 °C are conducted in flowing high-purity argon or a vacuum to prevent surface oxidation of SiC, which would alter the measured length and distort the CTE calculation.

What sample size and measurement uncertainty can be expected?

Standard bars of 3 mm × 3 mm × 10 mm or rods of 6 mm diameter × 20 mm are typical. With proper calibration against certified reference materials, the expanded uncertainty of the mean CTE is generally better than ±0.1 × 10⁻⁶ /K, or approximately ±2% relative.

How long does a complete measurement take?

A single heating and cooling cycle from room temperature to 1,000 °C at 5 K/min requires about 8 hours including stabilization and baseline correction. A full anisotropic characterization with repeat runs and XRD verification is typically completed within 3 to 5 working days.

In summary, the PC SiC single crystal thermal expansion coefficient measurement experiment provides certified, orientation-resolved CTE data that underpin epitaxy control, wafer handling, packaging reliability, and multiphysics simulation for silicon carbide power devices. Combining calibrated dilatometry with independent X-ray diffraction verification ensures that the reported expansion coefficients are both accurate and traceable.