Underfill Adhesive Coefficient of Thermal Expansion (CTE) Testing – Methods, Standards, and Applicat

Underfill adhesive coefficient of thermal expansion (CTE) testing is a critical thermomechanical characterization service for evaluating the dimensional stability of encapsulant materials used in flip-chip, BGA, and advanced semiconductor packaging. Accurate CTE measurement below and above the glass transition temperature helps manufacturers predict solder joint reliability, reduce delamination risk, and ensure long-term performance of electronic assemblies. This article provides a comprehensive technical overview of underfill CTE testing samples, test items, internationally recognized methods, typical applications, and frequently asked questions.

Technical Overview

Underfill adhesives are epoxy-based polymer materials dispensed or molded into the gap between a silicon die and its substrate in flip-chip, chip-scale package (CSP), and ball grid array (BGA) assemblies. Once cured, the underfill mechanically couples the chip to the substrate and redistributes thermomechanical stress away from the solder interconnects, dramatically improving thermal cycling fatigue life.

The coefficient of thermal expansion (CTE) is arguably the single most important thermomechanical property of an underfill adhesive. During operation, electronic assemblies undergo repeated temperature excursions, and each material in the stack-up—silicon die (~2.8 ppm/°C), solder joints (~21–25 ppm/°C), organic substrate (~14–17 ppm/°C), and cured underfill—expands at a different rate. The magnitude of the CTE mismatch directly determines the shear strain imposed on solder joints and the interfacial stress that can drive delamination or cracking. A well-formulated underfill typically exhibits a CTE of 20–40 ppm/°C below its glass transition temperature (Tg), striking a balance between matching the solder and maintaining stiffness.

Because polymeric underfills are viscoelastic materials, their CTE is not a single constant: the material expands at a low rate (α1) below Tg and at a substantially higher rate (α2) above Tg. Reliable CTE measurement therefore requires precise instrumentation, controlled heating rates, and well-defined data analysis windows, which is why manufacturers, packaging houses, and material suppliers rely on accredited third-party laboratories for qualified testing.

Testing Samples

Cured Bulk Underfill Specimens

Liquid or film-form underfill adhesives are cast into plaques or molds and fully cured according to the supplier-recommended cure schedule (for example, 150–165 °C for 30–120 minutes, followed by a post-cure if specified). Specimens are then machined or cut into rectangular prisms with flat, parallel, smooth end faces—typically 3–10 mm in length and 3–5 mm in cross-section—to suit thermomechanical analysis (TMA) in compression or expansion mode.

Cured No-Flow and Snap-Cure Underfills

No-flow underfills containing fluxing agents and fast-curing snap-cure formulations are prepared as bulk buttons or cured directly in standardized molds. Because these chemistries may cure partially during reflow, laboratories replicate the complete user process profile—reflow plus hold—so the specimen reflects the actual degree of cure achieved in production.

Capillary Underfill on Actual Assemblies

In some reliability investigations, CTE-relevant strain behavior is evaluated on actual flip-chip or BGA hardware to validate material model inputs for finite element simulation of package warpage and solder joint stress.

Sample Preparation and Conditioning Requirements

• Specimens must be free of voids, bubbles, cracks, and embedded filler segregation, all of which distort dimensional change measurements.
• End faces must be parallel within tight tolerances to ensure uniform contact with the TMA probe.
• Specimens are conditioned per standard practice (commonly 23 ± 2 °C and 50 ± 5% relative humidity) for a minimum of 24–40 hours before testing, unless a specific as-received or moisture-saturated condition is required.
• Moisture-sensitive grades may require desiccation or bake-out (for example, 125 °C for 24 hours) prior to high-temperature scanning to avoid moisture-induced artifacts near 100 °C.
• The degree of cure should be verified (e.g., by DSC residual enthalpy) whenever a new cure profile is used, since incomplete cure depresses Tg and inflates measured CTE.

Testing Items

Core thermomechanical parameters delivered in a typical underfill CTE test report include:

• Coefficient of linear thermal expansion below Tg (α1, ppm/°C), typically reported between 30–50 °C and the onset of the glass transition region.
• Coefficient of linear thermal expansion above Tg (α2, ppm/°C), typically reported between the end of the transition region and the maximum test temperature.
• Glass transition temperature (Tg) determined from the TMA expansion inflection, complementing DSC-based Tg where specified.
• Dimensional change (ΔL/L0, %) across the full scan range, useful for warpage modeling inputs.
• Linear expansion curve and tangent/Secant CTE as a function of temperature, provided in graphical form for simulation engineers.
• Optional comparative assessments: CTE after moisture preconditioning (85 °C/85% RH), after thermal aging, or across multiple cure schedules.

Testing Methods

Thermomechanical Analysis (TMA) – Primary Method

TMA is the industry-preferred technique for underfill CTE measurement. A specimen is placed on a quartz platform and contacted by a low-mass quartz probe under a small static force (typically 0.05–0.20 N). The furnace ramps at a controlled heating rate, commonly 3–5 °C/min (with 10 °C/min used in some protocols), over a temperature range such as −50 °C to 250 °C, while a high-resolution LVDT sensor tracks dimensional change to sub-micron resolution. CTE is calculated as (1/L0) × (dL/dT), with α1 and α2 extracted from regression over defined temperature windows. This approach conforms to ASTM E831 (Standard Test Method for Linear Thermal Expansion of Solid Materials by Thermomechanical Analysis) and ISO 11359-2 (Plastics — Thermomechanical analysis — Determination of coefficient of linear thermal expansion and glass transition temperature).

Dilatometry (ASTM D696)

Vitreo-silica dilatometry per ASTM D696 measures linear thermal expansion of plastics over a narrow, near-ambient range (−30 °C to +30 °C). It is less common for full underfill characterization because it cannot capture Tg or high-temperature behavior, but it remains a valid reference method for verifying low-temperature CTE values.

Differential Scanning Calorimetry (DSC) – Supporting Characterization

DSC per ISO 11357-2 or ASTM D3418 is routinely run alongside TMA to independently confirm Tg and verify degree of cure, ensuring the CTE data are interpreted against the correct transition boundaries.

Testing Conditions and Data Quality Controls

• Nitrogen purge (40–60 mL/min) is used to prevent oxidative degradation during high-temperature scans.
• Specimens are often pre-scanned once (thermal conditioning) and a second scan reported, or the first scan reported per agreed protocol; the approach is documented in the report.
• A minimum of two to three replicate specimens is tested, and repeatability typically falls within ±1–2 ppm/°C for α1.
• Calibration of the length-change sensor and temperature axis is verified with certified reference materials (e.g., aluminum and quartz standards) traceable to national metrology institutes.
• Relevant reference standards also include GB/T 36863 and related national plastic thermomechanical testing specifications where reporting to Chinese market frameworks is required.

Applications

Underfill CTE test data directly support design, qualification, and failure analysis activities across the electronics manufacturing chain:

• Flip-chip and 2.5D/3D advanced packaging: selecting underfills whose CTE minimizes solder bump shear strain during thermal cycling (e.g., −40 °C to +125 °C, per JEDEC JESD22-A104 profiles).
• BGA and CSP board-level reliability: validating reworkable and no-flow underfills before deployment in high-reliability products such as automotive ECUs, aerospace avionics, and telecommunications infrastructure.
• Material qualification and incoming inspection: comparing supplier lots or candidate formulations against datasheet CTE limits and internal specifications.
• Fan-out wafer-level packaging and chip-first/chip-last process development, where CTE influences wafer warpage and molding compound interaction.
• Finite element analysis (FEA) and lifetime prediction: supplying temperature-dependent CTE curves for ANSYS/ABAQUS thermo-mechanical models used to estimate solder joint fatigue life (e.g., Coffin–Manson based).
• Failure analysis: detecting CTE drift after thermal aging, moisture uptake, or incomplete cure as a root cause of delamination, popcorning, or underfill cracking.
• Automotive electronics per AEC-Q100/Q101-aligned qualification programs, where sustained high operating temperatures demand rigorous α1/α2 and Tg verification.

FAQ

What is a typical CTE target for underfill adhesives?

Most capillary underfills for solder interconnect protection are specified between 20 and 40 ppm/°C below Tg, which approximates the effective CTE of the solder/die/substrate system and keeps interfacial shear stress manageable. Above Tg, CTE commonly rises to 80–150 ppm/°C, which is why the operational temperature range of the device should always sit safely below the underfill Tg.

Why are two CTE values (α1 and α2) reported?

Cured epoxy underfills are amorphous crosslinked polymers. Below Tg the network is glassy and expansion is restricted (α1); above Tg the network gains segmental mobility and expands markedly faster (α2). Reporting both values, along with Tg, gives packaging engineers the complete expansion profile needed for simulation across the full temperature range.

How long does an underfill CTE test take?

A single TMA run covering −50 °C to 250 °C at 5 °C/min takes roughly 1.5–2 hours including equilibration. With specimen preparation, conditioning, replicate runs, and reporting, typical laboratory turnaround is 5–10 working days, depending on sample quantity and conditioning requirements.

Can CTE be measured on underfill cured inside an actual package?

Bulk cured specimens are strongly preferred for accurate material CTE because geometry, probe contact, and constraint effects complicate in-situ measurement. However, package-level techniques such as shadow moiré warpage measurement or strain gauge methods can validate the effective expansion behavior of the complete assembly against material data.

What sample size is required for TMA testing?

Most TMA expansion fixtures accept specimens 3–10 mm in height with a cross-section of about 3–5 mm square. Flat, parallel, void-free end faces are essential; laboratories can often prepare compliant specimens from customer-supplied cured plaques if bulk samples are unavailable.

How does incomplete cure or moisture affect CTE results?

Incomplete cure lowers Tg and generally increases α1, while absorbed moisture plasticizes the network and can create a measurement artifact near 100 °C. Reputable laboratories therefore verify cure state by DSC and apply documented conditioning or bake-out procedures before TMA scanning, and report all preparation details for full traceability.

Which standards govern underfill CTE testing reports?

The primary frameworks are ASTM E831 and ISO 11359-2 for TMA-based linear expansion, ASTM D696 for dilatometry, and ISO 11357-2/ASTM D3418 for supporting Tg determination, with JEDEC thermal cycling standards providing the qualification context in which the data are applied. Reports from accredited third-party laboratories should identify the applied standard, specimen geometry, heating rate, temperature windows used for α1/α2 regression, and calibration traceability.