Underfill Adhesive Curing Expansion Evaluation: Test Items, Methods, and Standards

Underfill adhesive curing expansion evaluation is a critical reliability assessment in advanced electronic packaging, quantifying how epoxy-based underfill materials expand, shrink, and stabilize during thermal cure and subsequent service. This technical article explains the key test items, sample preparation requirements, instrumentation, and applicable standards—such as ASTM E831, ISO 11359-2, and IPC TM-650—used by accredited third-party laboratories to characterize coefficient of thermal expansion, glass transition temperature, and cure-induced dimensional change. Understanding these parameters enables engineers to minimize warpage, solder joint stress, and delamination risk in flip-chip, BGA, CSP, and wafer-level assemblies.

Technical Overview

Underfill adhesives are epoxy-based encapsulants dispensed beneath flip-chip dies, ball grid arrays (BGA), and chip-scale packages (CSP) to redistribute thermomechanical stress away from fragile solder interconnects. During the curing process, these materials undergo complex dimensional changes driven by three simultaneous phenomena: thermal expansion as the assembly is heated to the cure temperature, chemical cure shrinkage as crosslinking reactions densify the polymer network, and post-gelation expansion as the material passes through its glass transition. The net dimensional behavior, collectively referred to as curing expansion, directly determines the residual stress state of the finished package.

Curing expansion evaluation is essential because of the severe coefficient of thermal expansion (CTE) mismatch in a typical package: silicon die at approximately 2.5 ppm/°C, solder joints at approximately 21–25 ppm/°C, organic substrates at 15–17 ppm/°C, and cured underfill at 25–45 ppm/°C below Tg. An underfill that expands excessively during cure or exhibits an uncontrolled CTE can induce die warpage, bump cracking, fillet cracking, and interfacial delamination. Accurate characterization of expansion behavior therefore underpins material qualification, process window definition, and package reliability prediction.

Key Engineering Concerns

• Warpage of the die and substrate induced by asymmetric cure shrinkage and expansion
• Compressive and shear stress transferred to solder bumps during and after cure
• Void formation and migration caused by volume change and volatile evolution
• Fillet geometry instability and edge delamination at the die passivation interface
• Long-term dimensional drift during reflow excursions and temperature cycling

Testing Samples

Representative sampling is fundamental to obtaining meaningful curing expansion data. Accredited third-party laboratories typically evaluate several categories of specimens depending on the stage of material development and the intended end use.

Bulk Cast Specimens

Liquid capillary underfills are cast into silicone or PTFE molds and cured according to the supplier-recommended profile (for example, 150 °C for 60 minutes or 165 °C for 30 minutes). Standard geometries include discs of 4–6 mm diameter and 2–3 mm thickness for thermomechanical analysis, bars of 5–10 mm length for dilatometry, and plaques for density and gravimetric measurements.

No-Flow and Snap-Cure Underfills

No-flow fluxing underfills, which cure during the solder reflow cycle, are cast and cured using a simulated reflow profile with peak temperatures of 220–260 °C. The evaluation must capture the rapid gelation kinetics and the associated expansion-shrinkage transitions that occur within seconds of reaching peak temperature.

Pre-Applied and Wafer-Level Underfill Films

Film-type and wafer-level underfills are laminated between prepared substrates or tested as free-standing films after B-stage cure. Specimens are conditioned at 23 ± 2 °C and 50 ± 5% relative humidity for a minimum of 24 hours prior to testing unless otherwise specified.

Assembled Package Coupons

Actual flip-chip or BGA assemblies are submitted to verify expansion behavior in situ. These coupons allow correlation between bulk material data and real package warpage measured by shadow moiré, digital image correlation, or profilometry.

Testing Items

A comprehensive curing expansion evaluation program covers the following measurable parameters:

• Coefficient of thermal expansion below the glass transition (alpha 1, ppm/°C)
• Coefficient of thermal expansion above the glass transition (alpha 2, ppm/°C)
• Glass transition temperature (Tg) by TMA, DSC, and DMA methods
• Linear and volumetric cure shrinkage (percent)
• Net curing expansion during the cure ramp and isothermal dwell
• Degree of cure and residual exotherm by differential scanning calorimetry
• Post-cure dimensional stability during a second thermal excursion
• Moisture-induced expansion after humidity preconditioning
• Volatile content and outgassing during cure
• Warpage of assembled coupons before and after cure

Testing Methods

Thermomechanical Analysis (TMA)

TMA is the primary technique for curing expansion evaluation. A cured disc specimen is heated at a controlled rate, typically 5–10 °C/min from room temperature to 250 °C, under a small static load while a dimension change probe records expansion. The slope of the dimension-temperature curve yields alpha 1 and alpha 2, and the inflection point identifies Tg. Testing follows ASTM E831 and ISO 11359-2, with CTE values in the glassy region commonly required to fall between 25 and 45 ppm/°C for capillary underfills.

Differential Scanning Calorimetry (DSC)

DSC testing per ASTM E1356, ASTM D3418, and ISO 11357-2 establishes Tg, residual cure exotherm, and degree of conversion. Comparing the first and second heating scans separates the contribution of post-curing from genuine thermal expansion, which is essential for interpreting TMA results on partially cured materials.

Push-Rod Dilatometry

For higher-precision bulk expansion measurement, bar specimens are evaluated with a vitreous silica dilatometer per ASTM E228 or ASTM D696. Dilatometry provides expanded temperature coverage and validates TMA-derived CTE values, particularly for materials used in automotive power modules that operate from -40 °C to 150 °C.

Density and Volumetric Shrinkage Methods

Volumetric cure shrinkage is determined by comparing the density of the liquid adhesive with the density of the fully cured solid using the Archimedes method per ASTM D792 or ISO 1183. Combined with TMA expansion data, this yields the net volume change the package experiences during processing.

In-Situ Cure Expansion Monitoring

Laser displacement sensing, embedded strain gauges, and dielectric analysis (DEA) track dimension change in real time while the specimen passes through the actual cure profile. This captures transient expansion peaks that post-cure TMA scans cannot resolve and supports cure kinetics modeling for fast-cure and snap-cure formulations.

Package-Level Verification

Assembled coupons undergo warpage measurement at room temperature, at cure temperature, and after JEDEC JESD22-A104 temperature cycling, with moisture preconditioning per JEDEC J-STD-020 where applicable. IPC-SM-817 and IPC TM-650 method 2.4.24 provide additional guidance for adhesive qualification and CTE measurement within electronics manufacturing contexts.

Applications

Curing expansion evaluation supports material selection and process release across a broad range of packaging platforms:

• Flip-chip underfill for high-performance CPUs, GPUs, and AI accelerators
• BGA and CSP corner and edge bonding in mobile and wearable devices
• Stacked-die and package-on-package memory modules
• Fan-out wafer-level packaging and pre-applied underfill films
• Automotive power modules and engine control units requiring AEC-Q-oriented material qualification
• High-density interconnect substrates and rigid-flex assemblies
• Semiconductor packaging material incoming inspection and supplier change qualification

Design teams use the resulting alpha 1, alpha 2, Tg, and shrinkage data as inputs for finite element warpage and stress simulations, while manufacturing engineers use them to define cure oven profiles, hold times, and rework windows.

FAQ

What is curing expansion in underfill adhesives?

Curing expansion is the net dimensional change an underfill undergoes as it is heated, gelled, and crosslinked during the cure cycle. It combines thermal expansion during ramp-up, chemical shrinkage during polymerization, and relaxation effects near the glass transition, and it determines the residual stress locked into the finished package.

What CTE values are considered acceptable for a cured underfill?

Most qualified capillary underfills exhibit alpha 1 between 25 and 45 ppm/°C and alpha 2 below approximately 150 ppm/°C. The ideal CTE is close to that of the solder joint, near 25 ppm/°C, to minimize shear strain at the interconnect during thermal cycling.

Which standards govern curing expansion testing?

CTE and Tg measurement are standardized in ASTM E831, ASTM E228, ASTM D696, and ISO 11359-2; Tg and cure conversion in ASTM E1356, ASTM D3418, and ISO 11357-2; density in ASTM D792 and ISO 1183; and package-level verification in JEDEC JESD22-A104, JEDEC J-STD-020, IPC-SM-817, and IPC TM-650.

How are liquid underfills prepared for expansion testing?

The liquid adhesive is degassed, cast into molds, and cured using the production cure profile. The cured specimens are then conditioned at 23 ± 2 °C and 50 ± 5% relative humidity before TMA, dilatometry, DSC, and density measurements are performed.

Why does moisture affect the expansion results?

Epoxy underfills absorb moisture that plasticizes the network, lowers the effective Tg, and generates additional swelling. Preconditioning per JEDEC J-STD-020 prior to measurement is therefore necessary when the package will experience reflow or humid service environments.

How long does a typical evaluation take?

A standard evaluation covering CTE, Tg, cure shrinkage, and degree of cure on bulk specimens is generally completed within five to ten working days at an accredited third-party laboratory, while package-level warpage and temperature cycling programs require additional time based on the specified cycle count.