Substrate coefficient of thermal expansion (CTE) measurement quantifies the dimensional stability of printed circuit board laminates, ceramic carriers, and flexible base films under thermal loading, and it is one of the most decisive material parameters for interconnection reliability. Accredited third-party laboratories perform substrate CTE determination in accordance with IPC-TM-650 2.4.24, ASTM E831, ASTM E228, and ISO 11359-2 using thermomechanical analysis (TMA), push-rod dilatometry, and non-contact optical techniques. This article presents the testable sample types, measured parameters, standard test procedures, typical reference values, and engineering applications of substrate CTE testing.
Technical Overview
The coefficient of thermal expansion (CTE) describes how much a substrate dimension changes per degree of temperature change. It is defined as α = (1/L₀) × (ΔL/ΔT), where L₀ is the initial specimen length (or thickness) at the reference temperature, and it is expressed in ppm/°C (10⁻⁶/°C). Because electronic assemblies stack materials with very different expansion behavior—copper at approximately 17 ppm/°C, silicon at 2.6 ppm/°C, and glass-fabric-reinforced epoxies at 12–17 ppm/°C in-plane—every thermal excursion generates interfacial stress proportional to the CTE mismatch.
For organic substrates, CTE is not a single number. Below the glass transition temperature (Tg), the laminate exhibits a relatively low expansion rate referred to as alpha 1 (α1). Above Tg, the polymer network softens and the expansion rate rises steeply to alpha 2 (α2), which for FR-4-type materials in the z-axis can reach 200–400 ppm/°C. During lead-free reflow, where peak temperatures of 245–260 °C are well above Tg, this elevated expansion drives plated-through-hole barrel cracking, inner-layer copper fatigue, and resin delamination. Accurate CTE determination is therefore a core qualification requirement for copper-clad laminates (CCL), prepreg, ceramic substrates, and flexible base films.
Substrate CTE testing is typically performed by accredited third-party laboratories using thermomechanical analysis in accordance with IPC-TM-650 method 2.4.24, ASTM E831, and ISO 11359-2, or by push-rod dilatometry in accordance with ASTM E228, supplemented by JIS C 6481, IEC 61189-3, and corresponding GB/T national standards for copper-clad laminate test methods.
Testing Samples
Rigid Organic Laminates (CCL and Prepreg)
FR-4, high-Tg FR-4 (Tg 170–180 °C), CEM-1/CEM-3 composite laminates, bismaleimide-triazine (BT), polyphenylene ether (PPE), cyanate ester, and polyimide glass laminates, as well as B-stage prepreg cured into plaques. Copper foil is normally removed by chemical etching so that the resin-glass system alone is characterized; copper-clad panels can also be measured in the as-delivered condition when composite expansion is of interest.
Ceramic Substrates
96% and 99.6% alumina (Al₂O₃), aluminum nitride (AlN), silicon nitride (Si₃N₄), zirconia-toughened ceramics, low-temperature co-fired ceramic (LTCC) tapes, and thick-film or thin-film metallized carriers used in power modules and RF packages.
Flexible and Semi-Flexible Substrates
Polyimide (PI) base films, adhesive and adhesiveless flexible copper-clad laminates (FCCL), PET films, and liquid crystal polymer (LCP) films for high-frequency flexible circuits.
Metal-Based and Specialty Substrates
Insulated metal substrates (IMS) with aluminum or copper base plates, metal-core boards, silicon wafer and glass core substrates for advanced packaging (including through-glass via designs), and build-up film substrates for BGA and flip-chip packages.
Specimens for z-axis TMA are typically cut to 6.4 mm × 6.4 mm (0.25 in × 0.25 in) squares with thickness between 0.3 and 3 mm, while dilatometer specimens require bars of 25–50 mm in length with flat, parallel faces. Prior to testing, samples are conditioned by baking at 105 °C for 2 h to eliminate absorbed moisture that would otherwise distort the expansion curve.
Testing Items
Standard substrate CTE test reports include the following parameters:
- In-plane (x/y) coefficient of thermal expansion below Tg (α1), ppm/°C
- In-plane CTE above Tg (α2), ppm/°C
- Z-axis (through-thickness) CTE below Tg (α1), ppm/°C
- Z-axis CTE above Tg (α2), ppm/°C
- Glass transition temperature (Tg) determined as the extrapolated intersection of the α1 and α2 tangent lines
- Total thickness expansion between 50 °C and 260 °C, expressed as a percentage
- Optional items: expansion hysteresis across thermal cycling, CTE of the copper-clad composite, and low-temperature (−55 to 25 °C) expansion behavior
Testing Methods
Thermomechanical Analysis (TMA): Z-Axis CTE and Tg
TMA per IPC-TM-650 2.4.24, ASTM E831, and ISO 11359-2 is the primary method for through-thickness expansion. A quartz or fused-silica expansion probe applies a small static force (typically a few grams, approximately 0.05 N) to the 6.4 mm square specimen resting on a flat platform. The furnace is purged with nitrogen or helium at 25–50 mL/min, and the specimen is heated from ambient temperature (or −50 °C for low-temperature studies) to 260–300 °C at 5 °C/min. The instrument records thickness change versus temperature; CTE is calculated from the slope of the linear segment roughly 20–30 °C below the transition region for α1 and 20–30 °C above it for α2, with Tg reported at the tangent intersection. Instruments are calibrated for temperature using indium and zinc melting-point standards and for expansion using certified sapphire references traceable to national metrology institutes.
Push-Rod Dilatometry: In-Plane CTE
Horizontal or vertical dilatometers operating per ASTM E228 and ISO 17562 measure the expansion of machined bars with a precision suitable for both organics and ceramics at temperatures up to and beyond 1000 °C. Because the measurement axis lies in the plane of the laminate, this method captures the x/y expansion governed by the glass fabric and ceramic microstructure, which differs markedly from the z-axis value of the same material. Many laminate qualification programs require both orientations to fully characterize anisotropic expansion.
Non-Contact and Strain-Based Techniques
Thin films below approximately 100 µm, large panels, and metallized specimens can be measured with high-temperature foil strain gauges, digital image correlation (DIC), or laser interferometry, which avoid probe-loading artifacts. For deposited films on wafers, the substrate-curvature method (Stoney equation) and X-ray diffraction sin²ψ analysis provide film CTE without fabricating freestanding specimens.
Test Procedure Summary
- Specimen preparation: cut coupons, etch copper if required, grind ends parallel, and mark orientation for x/y versus z-axis testing
- Conditioning: bake at 105 °C for 2 h, cool in a desiccator, and measure initial dimensions to within ±2.5 µm
- Calibration: verify force, temperature (indium/zinc), and expansion (sapphire) before each test campaign
- Measurement: apply probe load, purge with inert gas, and heat at 5 °C/min through the programmed range while recording dimensional change
- Data reduction: compute α1 and α2 slopes, tangent-intersection Tg, and total percentage expansion; report results to the nearest 1 ppm/°C
Applications
CTE data from substrate testing directly support design, qualification, and failure analysis across the electronics industry:
- Copper-clad laminate qualification against IPC-4101 slash-sheet and customer-specific maximum z-expansion limits, commonly 3.0–4.0% from 50 to 260 °C
- Plated-through-hole and microvia reliability assessment, since z-axis expansion above Tg is the dominant driver of copper barrel cracking during reflow and thermal cycling
- Package substrate selection, matching substrate CTE to silicon (2.6 ppm/°C), molding compounds, underfill, and solder to control warpage and joint stress in BGA, flip-chip, and fan-out packages
- Power module engineering using AlN (approximately 4.5 ppm/°C) or Si₃N₄ (approximately 3 ppm/°C) ceramic substrates to reduce solder-layer fatigue under thermal cycling
- High-layer-count board stack-up design and hybrid laminate selection for automotive, 5G/RF, aerospace, and high-density interconnect (HDI) products
- Failure analysis of delamination, blistering, and interconnect opens observed after reflow or thermal cycling, by comparing as-received and reflowed specimens
- Research and development of low-CTE resin systems, low-loss glass fabrics, and build-up dielectric films
FAQ
What is the difference between α1 and α2 CTE?
α1 is the expansion coefficient below the glass transition, where the polymer matrix is rigid and glass-fabric restraint dominates; α2 applies above Tg, where the softened resin expands much faster, particularly in the z-axis. Both values must be reported for organic substrates because reflow temperatures exceed Tg for most standard laminates.
Why does z-axis CTE matter more than x/y CTE for PCB reliability?
The glass fabric constrains in-plane expansion to roughly 12–17 ppm/°C, close to that of copper, while the z-axis is resin-dominated and can expand at 40–70 ppm/°C below Tg and several hundred ppm/°C above it. This differential expansion strains plated-hole copper and is the leading cause of barrel cracks and via failures during thermal excursions.
What are typical acceptance values for FR-4 substrates?
Common specification limits for FR-4 and high-Tg FR-4 include an x/y α1 of 12–17 ppm/°C, a z-axis α1 of 40–60 ppm/°C, a z-axis α2 of 200–350 ppm/°C, and total z-expansion of no more than 3.0–4.0% between 50 and 260 °C; high-reliability, lead-free-capable laminates typically target 2.5–3.0% or lower.
Which method should be chosen: TMA or dilatometry?
TMA (IPC-TM-650 2.4.24, ASTM E831, ISO 11359-2) is preferred for z-axis expansion, Tg, and small laminate coupons, while dilatometry (ASTM E228, ISO 17562) offers high precision for in-plane expansion of longer bars and for ceramics at elevated temperatures. Many qualification programs require both orientations to fully characterize the material.
What sample size and conditioning are required?
For z-axis TMA, a 6.4 mm × 6.4 mm coupon of 0.3–3 mm thickness is sufficient; dilatometry requires bars approximately 25–50 mm long. Copper may need to be etched away, and specimens should be baked at 105 °C for 2 h and stored in a desiccator before measurement to remove moisture.
Can CTE be measured below room temperature or after reflow simulation?
Yes. Sub-ambient expansion from −55 to 25 °C is measured for automotive and aerospace applications using cooled TMA or dilatometer furnaces, and specimens can be pre-conditioned through simulated reflow profiles or multiple thermal cycles to evaluate CTE drift and expansion hysteresis.