Gallium nitride (GaN) wafers are the foundation of next-generation power electronics, RF devices, and optoelectronics, and their thermal expansion coefficient (CTE) directly governs epitaxial stress, wafer bow, and long-term device reliability. Professional CTE evaluation of GaN wafers is typically performed by accredited third-party laboratories using thermomechanical analysis, dilatometry, and high-temperature X-ray diffraction in accordance with ASTM and ISO standards. This article explains the technical background, sample requirements, testing methods, and typical applications of GaN wafer thermal expansion coefficient evaluation.
Technical Overview
Gallium nitride is a wide-bandgap (approximately 3.4 eV) compound semiconductor with a wurtzite hexagonal crystal structure. Because this structure is anisotropic, GaN exhibits two principal linear thermal expansion coefficients: an in-plane coefficient (αa) along the a-axis and an out-of-plane coefficient (αc) along the c-axis. At room temperature, typical literature values are approximately 5.59 × 10⁻⁶ K⁻¹ for αa and 3.17 × 10⁻⁶ K⁻¹ for αc, and both values increase measurably with temperature.
Accurate CTE data is critical because the majority of GaN wafers are heteroepitaxial structures grown on foreign substrates such as silicon, silicon carbide, or sapphire. The CTE mismatch between the GaN layer and the substrate generates thermal stress during post-growth cooling and during subsequent device fabrication and operation. Excessive mismatch leads to wafer bow, warp, micro-cracking, delamination, and degraded carrier mobility in the active channel. Reliable CTE measurement therefore supports epitaxial recipe development, substrate selection, thermal-mechanical simulation, and packaging design.
Because bulk GaN wafers are small, thin, hard, and anisotropic, their CTE evaluation requires specialized instrumentation and careful sample preparation. Accredited third-party laboratories typically apply thermomechanical analysis (TMA), push-rod dilatometry, and high-temperature X-ray diffraction (HT-XRD) to obtain traceable results with expanded uncertainties on the order of a few percent.
Testing Samples
Free-Standing GaN Wafers
Self-supported GaN wafers produced by hydride vapor phase epitaxy (HVPE) or ammonothermal growth, commonly in 2-inch, 4-inch, or 6-inch formats with thicknesses from 300 μm to over 1 mm. These samples allow direct measurement of the intrinsic CTE of GaN in both the in-plane and out-of-plane directions.
GaN-on-Silicon Epiwafers
GaN heterostructures grown on Si(111) substrates for power devices. The composite CTE behavior of the stack is dominated by the thick silicon substrate above approximately 200 °C, so testing focuses on characterizing the effective CTE of the finished wafer and the stress state of the GaN film.
GaN-on-SiC Epiwafers
RF-grade GaN layers on semi-insulating SiC substrates. The small but non-negligible CTE mismatch between GaN and SiC is a key reliability factor for high-power amplifiers, making precise CTE characterization of both materials essential.
GaN-on-Sapphire Epiwafers
LED and power structures on sapphire substrates, where the relatively large CTE mismatch requires careful stress management during growth and thermal processing.
Sample Preparation Requirements
- Cut specimens of 3–10 mm in length with two parallel, flat end faces for dilatometric measurement.
- Thickness of at least 0.3–0.5 mm for reliable TMA push-rod contact.
- Remove grinding damage by chemical-mechanical polishing to avoid spurious creep or microcrack effects.
- Clean surfaces with solvent and deionized water, then dry with nitrogen before testing.
- Document crystallographic orientation (a-axis or c-axis measurement direction) and wafer dopant or carrier concentration.
Testing Items
A comprehensive CTE evaluation program for GaN wafers typically includes the following measurable parameters:
- Mean (secant) linear CTE over defined temperature intervals, such as 25–200 °C, 25–500 °C, and 25–1000 °C.
- Instantaneous (tangent) CTE as a continuous function of temperature.
- In-plane expansion coefficient αa for wurtzite GaN along the a-axis.
- Out-of-plane expansion coefficient αc along the c-axis, quantifying crystallographic anisotropy.
- Effective through-thickness CTE of composite epiwafer stacks.
- Thermal cycling stability, comparing the CTE of the first and subsequent heating cycles to reveal relaxation or crack initiation.
- Repeatability and expanded measurement uncertainty (typically stated at 95% confidence, k = 2).
Testing Methods
Thermomechanical Analysis (TMA)
TMA is the most widely used method, performed in accordance with ASTM E831 or the ISO 11359 series. A quartz or alumina stage supports the specimen while a low-force probe records dimensional change during heating at 3–5 °C/min under nitrogen or argon. The mean CTE is calculated from α = ΔL / (L₀ × ΔT), where L₀ is the initial specimen length. High-purity corundum or fused silica reference materials are run to verify system calibration.
Push-Rod Dilatometry
Horizontal or vertical dilatometers following ASTM E228 provide excellent accuracy for specimens that can be prepared at least several millimeters long. A differential configuration with a matched reference holder suppresses apparatus expansion, achieving uncertainties approaching 1 × 10⁻⁷ K⁻¹ for well-conditioned samples.
High-Temperature X-Ray Diffraction (HT-XRD)
Because GaN is strongly anisotropic and many test pieces are thin, HT-XRD is often the preferred technique. Lattice parameters a(T) and c(T) are measured at stepped temperatures in a controlled atmosphere or vacuum, and the CTE values are derived from αa = (1/a₀)(da/dT) and αc = (1/c₀)(dc/dT). This method directly resolves both principal directions and can be applied to epitaxial films on their native substrates.
Laser Interferometry
For the highest precision, optical interferometric dilatometry per ASTM E289 measures displacement through fringe counting, with resolution in the nanometer range. It is used mainly as a reference technique for certifying material standards or resolving small anisotropic differences.
General Test Conditions
- Temperature range: room temperature to 600 °C for device-level studies, or up to 1000 °C for growth-simulation studies.
- Atmosphere: high-purity nitrogen, argon, or vacuum to prevent oxidation of GaN surfaces at elevated temperature.
- Pre-test conditioning: bake at 110–150 °C to remove absorbed moisture.
- At least two heating cycles to distinguish reversible expansion from irreversible structural changes.
- Calibration against certified reference materials traceable to national metrology institutes.
Applications
CTE evaluation results for GaN wafers are applied throughout the semiconductor value chain:
- Epitaxial growth optimization: selecting substrate materials and designing buffer layers that minimize thermal mismatch stress during cooldown from growth temperature.
- Wafer geometry control: predicting and correcting bow and warp introduced by CTE mismatch in HVPE, MOCVD, and MBE processes.
- Power electronics reliability: assessing thermal cycling endurance of GaN-on-Si HEMTs in inverters, fast chargers, and motor drives.
- RF and microwave devices: validating GaN-on-SiC material stability for radar, satellite communication, and 5G base-station amplifiers.
- Optoelectronics: ensuring mechanical integrity of LED epiwafers through die fabrication and packaging reflow steps.
- Packaging engineering: choosing die-attach materials, substrates, and molding compounds with matched expansion behavior to reduce solder fatigue and delamination.
- Finite element analysis: supplying accurate, temperature-dependent CTE curves as boundary conditions for thermal-mechanical simulation models.
FAQ
What are the typical CTE values of GaN compared with common substrates?
At room temperature, GaN expands at roughly 5.6 × 10⁻⁶ K⁻¹ in-plane and 3.2 × 10⁻⁶ K⁻¹ out-of-plane. Comparable values are approximately 4.5 × 10⁻⁶ K⁻¹ for SiC, 7.5–8.0 × 10⁻⁶ K⁻¹ for sapphire (in-plane), and 2.6–4.4 × 10⁻⁶ K⁻¹ for silicon depending on temperature, which explains why stress management strategies differ for each substrate system.
Why must GaN be measured in two directions?
GaN crystallizes in the hexagonal wurtzite structure, so its expansion is anisotropic. Reporting a single isotropic value understates in-plane stress and overstates out-of-plane expansion, leading to inaccurate warpage predictions in both growth simulation and packaging design.
What temperature range should be tested?
The range should match the thermal history of the wafer. Device-level studies usually cover 25–250 °C, reliability qualification extends to 300–500 °C, and growth-related studies may extend to 1000 °C or higher under protective atmosphere.
What is the difference between mean and instantaneous CTE?
The mean (secant) CTE is referenced to the starting temperature and describes total expansion over an interval, which is convenient for engineering calculations. The instantaneous (tangent) CTE is the derivative at each temperature and is required for accurate nonlinear finite element modeling.
How much sample is required for testing?
TMA and dilatometry generally require a strip of 3–10 mm length and at least 0.3 mm thickness with parallel faces. When wafer dimensions are too small, HT-XRD can determine both principal CTE values from pieces as thin as the epitaxial layer itself.
What measurement uncertainty can be expected?
With calibrated instruments and certified reference materials, dilatometric methods typically achieve expanded uncertainties of 2–5% for GaN-class materials, while interferometric methods can reach better than 1% under controlled laboratory conditions.
Which standards govern the test procedures?
The principal references are ASTM E831 for TMA, ASTM E228 for push-rod dilatometry, ASTM E289 for interferometric dilatometry, and the ISO 11359 series for thermomechanical analysis, supplemented by internal HT-XRD procedures validated against these frameworks.
