IC Substrate Linear Expansion Testing: Coefficient of Thermal Expansion (CTE) Measurement Guide

IC substrate linear expansion testing measures the coefficient of thermal expansion (CTE) of packaging substrates, a critical parameter that determines solder joint reliability, warpage behavior, and long-term durability of semiconductor packages. This guide covers the technical principles, sample requirements, test items, measurement methods, industry standards, and typical applications of CTE determination for IC substrates.

Technical Overview

The coefficient of linear thermal expansion (CTE, symbol α) quantifies the dimensional change of a material per unit length per degree of temperature change, expressed in ppm/°C. For IC substrates, linear expansion testing is one of the most important thermo-mechanical characterization procedures because the substrate acts as the mechanical and electrical interface between the silicon die and the printed circuit board. Since silicon has a CTE of approximately 2.6 ppm/°C while standard PCB material exhibits 14–17 ppm/°C, the substrate must bridge a significant thermal expansion mismatch that would otherwise concentrate stress at solder joints and die-attach interfaces.

IC substrates are inherently anisotropic composite structures consisting of resin, glass cloth or build-up dielectric layers, and copper circuitry. In-plane CTE (X/Y direction) is typically constrained by the glass fabric and ranges from 8 to 17 ppm/°C below the glass transition temperature, while the out-of-plane CTE (Z-axis, thickness direction) can reach 40–70 ppm/°C below Tg and exceed 150–250 ppm/°C above Tg. Accurate measurement of linear expansion in each direction, both below and above Tg, is therefore essential for package design, reliability prediction, and incoming material qualification.

Excessive or mismatched linear expansion leads to well-documented failure modes including solder joint fatigue during thermal cycling, interfacial delamination between the core and build-up layers, copper trace cracking, and package warpage that compromises assembly yield. Linear expansion testing allows engineers to select substrate materials whose CTE profile is compatible with the die, underfill, molding compound, and board, minimizing thermomechanical stress across the entire assembly.

Testing Samples

Rigid Organic Substrates

Rigid IC substrates based on bismaleimide triazine (BT) resin, epoxy/glass composites, and epoxy build-up films (such as ABF-type dielectrics) are the most commonly tested sample category. Specimens are typically cut as rectangular strips from production panels or dedicated coupons, with dimensions of approximately 3–6.35 mm in width and 10–25 mm in length. The sampling location and orientation relative to the glass cloth warp and fill directions must be documented because in-plane CTE differs between directions.

Flexible and Ceramic Substrates

Flexible substrates based on polyimide film require careful specimen handling due to their low bending stiffness, and their CTE is strongly influenced by the presence and layout of copper layers and adhesive systems. Ceramic substrates such as alumina (Al₂O₃, ~6.7 ppm/°C) and aluminum nitride (AlN, ~4.5 ppm/°C) exhibit low, nearly isotropic expansion and are tested when hermetic or high-power packaging compatibility must be verified.

Material Forms and Sample Preparation

Test specimens may be taken from core laminate sheets, build-up dielectric films, completed multilayer substrates, or copper-clad panels. The following preparation requirements apply:

• Specimen edges must be cut cleanly without delamination, burrs, or micro-cracks that could distort the expansion signal.
• Surfaces should be flat and parallel; grinding may be required for warped coupons so the probe seats properly.
• Moisture conditioning is mandatory—samples are typically baked at 105 °C for 1–2 hours and stored in a desiccator before testing, since absorbed moisture suppresses the measured Tg and alters the apparent expansion curve.
• Sample identification must record orientation (warp/fill/Z-axis), layer count, copper content, and thickness.
• A minimum of three specimens per direction is recommended to establish statistical validity.

Testing Items

Linear expansion testing of IC substrates generates a dimension-versus-temperature curve from which several quantitative results are extracted:

• CTE α1 (below Tg): The linear expansion coefficient measured below the glass transition region, typically evaluated between 30 °C and (Tg − 20 °C). This value governs thermomechanical stress during normal device operation.

• CTE α2 (above Tg): The expansion coefficient above Tg, evaluated after the transition region. Resin-dominated behavior causes a steep increase, and α2 is a key indicator of z-axis reliability risks during reflow.

• Glass transition temperature (Tg): Determined from the inflection point of the expansion curve, reported together with the CTE values.

• In-plane CTE (X and Y directions): Measured on edge-oriented specimens; critical for solder joint strain and die stress calculations.

• Out-of-plane CTE (Z-axis): Measured through the thickness; critical for plated via integrity and interfacial delamination assessment.

• Percentage of total expansion: The cumulative dimensional change over a defined range (for example, 30–260 °C), frequently specified in material datasheets.

Testing Methods

Thermomechanical Analysis (TMA)

TMA is the industry-standard method for IC substrate linear expansion measurement, performed according to IPC-TM-650 method 2.4.24 (and 2.4.24.1 for Z-axis), ASTM E831, and ISO 11359-2. A specimen is placed on a quartz stage and contacted by a low-mass expansion probe applying a static force of approximately 0.05–0.3 N. The furnace heats the specimen at a controlled rate, and a linear variable displacement transducer records dimensional change with sub-micron resolution.

Typical test conditions include:

• Temperature range: 30 °C to 260 °C (extended to 300 °C for reflow simulation)
• Heating rate: 5 °C/min per IPC-TM-650 (10 °C/min permitted by ASTM E831)
• Atmosphere: high-purity nitrogen purge at 20–50 mL/min to prevent oxidation
• Probe type: macro expansion probe (flat tip) for through-thickness measurement; special edge fixtures for in-plane specimens
• Calibration: temperature calibration with indium and zinc standards; displacement calibration with a certified reference material such as fused silica or aluminum

Push-Rod Dilatometry

Horizontal or vertical push-rod dilatometers per ASTM E228 accommodate larger specimens and deliver excellent accuracy for low-expansion materials. Dilatometry is often preferred for ceramic substrates and for verifying TMA results, although it requires longer specimen lengths and offers slower throughput.

Optical and Non-Contact Methods

Laser interferometry, digital image correlation (DIC), and shadow Moiré techniques measure in-plane expansion and warpage-related deformation without contact forces. These methods are valuable for thin flexible substrates, full-field strain mapping, and correlating free expansion behavior with package-level warpage measured per JEDEC JESD22-B112.

Data Evaluation

CTE is computed as α = ΔL / (L₀ × ΔT), where ΔL is the dimensional change, L₀ is the original length (or thickness), and ΔT is the temperature interval. Regression lines are fitted to the linear segments of the expansion curve below and above the transition, and Tg is reported as the intersection or the point of maximum slope change. Test reports should include the raw expansion curve, both CTE values, Tg, test standard, specimen orientation, and conditioning history.

Applications

Linear expansion data for IC substrates directly supports multiple engineering and quality functions across the semiconductor supply chain:

• Package design and material selection: matching substrate CTE with silicon die, underfill, mold compound, and PCB laminate to minimize thermomechanical stress in flip-chip, FCBGA, BGA, and chip-scale package architectures.

• Reliability qualification: input CTE values into finite element models and correlate with temperature cycling tests per JEDEC JESD22-A104 (−40 °C to +125 °C) to predict solder joint fatigue life and interconnect durability.

• Warpage control: Z-axis expansion and Tg data feed warpage simulation and process window definition for surface mount assembly and reflow.

• Incoming inspection and supplier qualification: PCB and package assemblers verify that substrate lots conform to datasheet CTE limits before production release.

• Manufacturing process control: substrate manufacturers monitor lamination and curing consistency, resin batch variation, and moisture uptake effects through routine expansion testing.

• Failure analysis: abnormal α2 values or expansion curve discontinuities reveal partial curing, degradation, or latent delamination in returned units.

FAQ

What CTE values are typical for IC substrates? Common rigid organic substrates exhibit in-plane CTE of 8–17 ppm/°C and Z-axis CTE of 40–70 ppm/°C below Tg, rising above 150 ppm/°C beyond Tg. Ceramic substrates range from about 4.5 to 7 ppm/°C depending on composition.

Why are two CTE values (α1 and α2) reported? Organic substrate resins soften above the glass transition temperature, causing a sharp increase in expansion rate. Reporting the coefficient both below and above Tg fully characterizes behavior across operating and reflow temperature ranges.

Which standards govern the test? IPC-TM-650 2.4.24 is the primary industry method for substrate CTE and Tg by TMA, complemented by ASTM E831, ASTM E228 (dilatometry), and ISO 11359-2 for plastics thermo-mechanical analysis.

How do in-plane and Z-axis measurements differ? In-plane specimens are oriented so the probe measures along the panel surface, capturing glass-cloth-constrained expansion relevant to solder joint strain. Z-axis testing measures through the thickness, where resin expansion dominates and plated via reliability is the concern.

What accuracy can be expected? A properly calibrated TMA achieves displacement resolution better than 0.1 µm; repeatability for in-plane CTE is typically within ±1 ppm/°C for values below 20 ppm/°C. Interlaboratory reproducibility improves significantly when heating rate, conditioning, and orientation are standardized.

Is moisture conditioning necessary? Yes. Absorbed moisture plasticizes the resin, lowering the apparent Tg and distorting the expansion curve. Baking specimens at 105 °C followed by desiccator storage ensures reproducible, comparable results.