Determination of the Coefficient of Thermal Expansion of Resin Fiberglass Boards

The coefficient of thermal expansion (CTE) is one of the most critical thermal-mechanical properties of resin fiberglass boards, directly governing the dimensional stability, plated-through-hole reliability, and warpage behavior of printed circuit boards and electronic packages. This article explains how CTE—including α1 below Tg, α2 above Tg, and both in-plane and z-axis values—is determined for glass-fabric-reinforced resin laminates such as FR-4 using thermomechanical analysis (TMA) and dilatometry in accordance with IPC-TM-650, ASTM, and ISO standards, covering specimen requirements, test procedures, typical results, and engineering applications.

Technical Overview

Resin fiberglass boards are composite laminates produced by impregnating woven electrical-grade glass fabric (typically E-glass) with a thermosetting resin system—most commonly epoxy, but also polyimide, bismaleimide-triazine (BT), phenolic, or modified high-Tg epoxies—and curing the stack under heat and pressure. Grades such as FR-4, G10, CEM-3, and high-Tg laminates form the backbone of the printed circuit board (PCB) industry and are also used as electrical insulation and light structural components. Because the laminate combines a low-expansion glass reinforcement with a comparatively high-expansion polymer matrix, its thermal expansion behavior is both anisotropic and strongly dependent on temperature.

The coefficient of thermal expansion (CTE) quantifies the fractional change in length per unit temperature change and is expressed in ppm/°C. It is defined as α = (1/L₀) × (ΔL/ΔT), where L₀ is the initial specimen length at the reference temperature. For resin fiberglass laminates, three characteristics are essential:

• Anisotropy: in-plane (x/y) expansion is constrained by the woven glass fabric, typically 10–17 ppm/°C, while through-thickness (z-axis) expansion is resin-dominated and considerably larger.
• Temperature dependence through the glass transition: below Tg the material exhibits a lower expansion coefficient α1; above Tg the matrix softens and the expansion coefficient α2 rises sharply, often by a factor of 4–8 in the z-axis.
• Directional dependence within the plane: warp and fill directions of the glass cloth may differ by 1–3 ppm/°C due to yarn crimp and fabric construction.

Accurate CTE determination is vital because excessive z-axis expansion during reflow soldering or thermal cycling stresses plated through-holes and vias, causing barrel cracking, inner-layer separation, and pad lifting. In-plane CTE mismatch with silicon (about 2.6 ppm/°C) or ceramic components drives solder-joint fatigue and package warpage. Product specifications such as IPC-4101 slash sheets therefore impose explicit CTE and total-expansion limits on qualified laminate grades, making CTE measurement a routine element of laminate qualification, incoming inspection, and failure analysis—whether performed in-house or by accredited third-party laboratories.

Testing Samples

FR-4 Epoxy Glass Laminates

Standard FR-4 copper-clad or bare laminates in thicknesses from about 0.2 mm to 3.2 mm represent the largest share of CTE testing volume. Specimens are usually cut with copper foil removed by chemical etching so that the resin-glass system itself is characterized, unless the clad laminate behavior is specifically required.

High-Tg and Lead-Free Compatible Laminates

Modified epoxy systems with TMA Tg values of 150–180 °C, often mineral-filled to reduce expansion, are tested over an extended temperature range (commonly to 260 °C or higher) to verify compliance with lead-free reflow requirements.

Polyimide, BT, and Specialty Laminates

High-temperature laminates for aerospace, automotive, and IC-substrate applications are measured to confirm low in-plane CTE and reduced z-axis expansion for reliability in aggressive thermal environments.

Composite Grades (CEM-1, CEM-3)

Laminates with paper or glass core constructions are tested to verify through-thickness expansion behavior, which differs from homogeneous FR-4 due to the layered structure.

Cured Prepreg Specimens

B-staged prepreg sheets can be fully cured in a plaque mold and subsequently tested to characterize the resin-glass system before laminate production.

Testing Items

A complete CTE characterization of a resin fiberglass board typically includes the following measured parameters:

• Z-axis (through-thickness) CTE below Tg (α1), typically reported over a defined window such as 50 °C to (Tg − 20 °C).
• Z-axis CTE above Tg (α2), reflecting the softened resin behavior.
• Total z-axis expansion (%) over a specified interval, commonly 50–260 °C, as required by many IPC-4101 slash sheets.
• In-plane (x and y) CTE below and above Tg, measured on strips cut along warp and fill directions.
• Glass transition temperature (Tg), determined simultaneously from the inflection point of the TMA expansion curve.
• Optional derived values: expansion at reflow peak temperature, residual cure shrinkage, and post-Tg expansion onset.

Testing Methods

Specimen Preparation and Conditioning

Correct specimen preparation is decisive for repeatable results. Standard practice per IPC-TM-650 2.4.24 and ASTM E831 includes:

• Z-axis specimens: square pieces of at least 6.4 mm × 6.4 mm, with the full laminate thickness (typically 0.8–3.2 mm) and both faces machined flat, parallel, and free of burrs.
• In-plane specimens: strips commonly 10–25 mm long and 3–7 mm wide, cut along the warp and fill axes and marked for orientation.
• Copper foil removed by etching unless the clad condition is explicitly required.
• Pre-drying (e.g., 105 °C for 1–2 h) followed by cooling in a desiccator to remove absorbed moisture, since moisture plasticizes epoxy and inflates apparent α values.
• Equilibration at 23 ± 2 °C and 50 ± 5% relative humidity prior to measurement.

Thermomechanical Analysis (TMA) — Primary Method

TMA is the industry-standard technique for resin fiberglass boards. The specimen is placed on a quartz platform and contacted by a low-force expansion probe (typically below 0.1 N); a linear variable displacement transducer records dimensional change as the furnace is heated. Typical conditions under IPC-TM-650 2.4.24, ASTM E831, and ISO 11359-2 are:

• Temperature program from ambient (or 25–50 °C start) to 260 °C for standard FR-4, or to 300 °C and beyond for high-Tg systems.
• Heating rate of 5 °C/min under a nitrogen purge to minimize oxidation of cut edges and probe drift.
• Calibration with a certified reference material (e.g., aluminum or quartz) for temperature and length-change accuracy.
• α1 and α2 obtained by linear regression on the pre- and post-transition segments of the expansion curve; Tg taken at the intersection of the extrapolated segments or the inflection point.

Push-Rod Dilatometry

For in-plane CTE and higher-precision measurements, fused-silica push-rod dilatometers per ASTM E228 (up to ~1000 °C capability) or quartz dilatometers per ASTM D696 (narrow range near room temperature) are used. Longer specimens improve resolution, making dilatometry well suited to warp/fill anisotropy studies and to low-expansion, highly filled laminate grades. Optical and interferometric methods (e.g., ASTM E289 type approaches) can be applied when very low CTE values must be resolved.

Data Reduction and Reporting

The mean CTE over any temperature interval is the slope of the normalized expansion curve (ΔL/L₀ versus T) obtained by least-squares regression. A compliant test report should state the material and grade, specimen orientation and dimensions, etching condition, conditioning history, instrument and method (TMA or dilatometer), heating rate and atmosphere, the regression windows used for α1 and α2, Tg, total expansion over the specified interval, and the number of replicates (typically three to five per condition).

Typical Results for Common Laminates

• Standard FR-4: in-plane α1 ≈ 12–17 ppm/°C; z-axis α1 ≈ 45–60 ppm/°C; z-axis α2 ≈ 200–350 ppm/°C.
• High-Tg, filled epoxies: in-plane α1 ≈ 10–14 ppm/°C; z-axis total expansion (50–260 °C) commonly ≤ 3.0–3.5%.
• Polyimide/glass: in-plane α1 ≈ 12–14 ppm/°C; z-axis α1 ≈ 35–50 ppm/°C with excellent post-Tg stability.
• BT/glass for substrates: in-plane α1 ≈ 12–15 ppm/°C with low total z-expansion.

Factors Affecting Accuracy

• Moisture in the resin matrix, which raises apparent expansion and depresses Tg.
• Residual cure and thermal history; a first heating scan may show cure-related shrinkage superimposed on expansion.
• Probe force and specimen surface parallelism, which influence contact and apparent length change.
• Heating rate discrepancies between laboratories; interlaboratory comparisons require matched programs.
• Resin content and glass style, which shift CTE between laminate lots and thicknesses.

Applications

CTE data for resin fiberglass boards underpin nearly every dimensioning and reliability decision in electronics manufacturing:

• Laminate qualification and incoming inspection against IPC-4101 slash sheet limits, supported by IEC 61189 series and JIS C 6481 test frameworks.
• Plated-through-hole and via reliability engineering, where z-axis expansion during reflow governs barrel stress and crack initiation.
• IC package substrate selection and CTE matching with silicon dies, ceramic carriers, and molding compounds to control warpage and solder-joint fatigue.
• Multilayer PCB stack-up design and warpage simulation, requiring accurate x, y, and z expansion inputs.
• Automotive and aerospace electronics subjected to thermal cycling from −40 °C to +150 °C or beyond.
• Quality control of filled, low-CTE laminate development programs and failure analysis of field returns.
• Structural uses of G10/FR-4 sheets—insulators, gears, standoff posts—where thermal clearance must be designed in.

FAQ

What is the difference between α1 and α2 in CTE testing?

α1 is the coefficient of thermal expansion below the glass transition temperature, where the resin matrix is rigid and expansion is comparatively low. α2 is the expansion coefficient above Tg, where the softened matrix expands much faster. For FR-4, z-axis α2 is typically several times larger than α1, which is why both values are reported.

Why is the z-axis CTE of resin fiberglass boards so much higher than the in-plane CTE?

The woven glass fabric carries load and constrains expansion in the x and y directions, while the z-axis contains only resin and glass thickness with no continuous fiber restraint, so through-thickness expansion is resin-dominated and much larger.

Which standards govern the measurement?

The dominant method for PCB laminates is IPC-TM-650 method 2.4.24 using TMA; equivalent generic methods include ASTM E831 and ISO 11359-2 for TMA, ASTM E228 for dilatometry, and ASTM D696 for near-room-temperature plastics. Acceptance limits are defined in IPC-4101 slash sheets, with related frameworks in IEC 61189 and JIS C 6481.

What specimen size is required for testing?

For z-axis TMA testing, a square of at least 6.4 mm × 6.4 mm at the full laminate thickness (about 0.8–3.2 mm) is standard. In-plane testing uses longer strips, commonly 10–25 mm, cut along the warp and fill directions. Three to five replicates are recommended for statistical confidence.

Should copper foil be removed before testing?

Yes, in most cases. Standard practice is to etch off the copper so that the expansion of the resin-glass laminate itself is measured; copper, with a CTE of about 17 ppm/°C, would otherwise bias the result. Testing with foil intact is only performed when the clad laminate behavior is specifically of interest.

What temperature range and heating rate are used?

Typical TMA programs run from ambient or 50 °C to 260 °C at 5 °C/min under nitrogen for FR-4 grades; high-Tg materials may be taken to 300 °C or higher to fully capture α2. Total z-expansion over 50–260 °C is the key qualification metric for lead-free compatibility.

What are typical acceptance values?

While limits vary by slash sheet and grade, standard FR-4 laminates are commonly specified with z-axis total expansion (50–260 °C) around 4.0% maximum, with premium filled, high-Tg systems at approximately 3.0% or lower, and in-plane α1 in the 10–17 ppm/°C range. The governing laminate specification should always be consulted.

Can Tg be determined in the same test?

Yes. The TMA expansion curve exhibits a clear slope change at the glass transition, so Tg (reported as TMA Tg) is obtained simultaneously with α1 and α2 from a single scan, making TMA an efficient multi-parameter measurement.

Where can this testing be performed?

Laminate manufacturers, PCB fabricators, and accredited third-party laboratories with calibrated TMA and dilatometry capability all routinely perform CTE determination. For compliance-related results, laboratories should hold ISO/IEC 17025 accreditation and report full test conditions alongside the measured values.