Thermal Expansion Testing and Analysis of Underfill Adhesives: CTE and Tg Characterization for Advan

Thermal expansion testing and analysis of underfill adhesives is a critical reliability evaluation that measures the coefficient of thermal expansion (CTE) and glass transition temperature (Tg) of encapsulant materials used in flip-chip, BGA, and wafer-level packages. Because mismatches in thermal expansion between the silicon die, solder interconnects, and organic substrate drive solder joint fatigue and delamination, accurate CTE characterization using thermomechanical analysis (TMA), dilatometry, and dynamic mechanical analysis (DMA) is essential for material qualification and failure prevention. This article explains the testing samples, measured parameters, standard methods, and typical applications of underfill adhesive thermal expansion analysis.

Technical Overview

Underfill adhesives are epoxy-based or epoxy-modified polymeric encapsulants dispensed into the gap between a silicon die and an organic laminate substrate in flip-chip, chip-scale package (CSP), ball grid array (BGA), and wafer-level package (WLP) assemblies. Once cured, the underfill mechanically couples the die to the substrate and redistributes thermomechanical stress away from the solder joints, typically improving solder joint fatigue life by a factor of five to ten under thermal cycling.

The effectiveness of an underfill depends almost entirely on how its thermal expansion behavior matches the surrounding materials. Silicon expands at approximately 2.6 ppm/°C, eutectic and SAC solder alloys at roughly 20–25 ppm/°C, and organic substrates at 14–17 ppm/°C in the in-plane direction. During temperature excursions, this mismatch generates shear strain at every die corner. An underfill with a well-controlled coefficient of thermal expansion (CTE), generally 25–45 ppm/°C below Tg, absorbs and redistributes this strain; an underfill with excessive or unstable CTE accelerates joint cracking, fillet delamination, and die backside stress.

Thermal expansion of a cured underfill is not a single constant. Below the glass transition temperature (Tg), the crosslinked network is in a glassy state and exhibits a low expansion rate, defined as alpha-1 (α1). Above Tg, molecular mobility increases sharply and expansion typically rises two to four times, defined as alpha-2 (α2). If Tg falls within the operating or reflow temperature range of the device, the localized expansion step change can cause sudden loss of adhesion or solder extrusion. Consequently, thermal expansion testing always evaluates CTE and Tg together, following recognized methods from ISO, ASTM, IPC, JEDEC, and GB standard systems as performed by accredited third-party laboratories and in-house reliability teams.

Testing Samples

Cured Bulk Specimens from Liquid Underfill

Liquid capillary or no-flow underfill adhesives are cast into flat molds or between release films, degassed, and cured according to the supplier-recommended schedule (for example, 150 °C for 30 to 60 minutes). After post-cure conditioning, free-standing strips are cut and surface-ground to obtain parallel, flat faces suitable for thermomechanical probing.

Pre-Cured Commercial Underfill Compounds

Solid underfill films, wafer-applied underfill coatings, and molded underfill compounds are tested as molded bars, pellets, or die-cut plaques. Because these materials are already B-staged or fully cured, specimen preparation focuses on achieving representative geometry and consistent thermal history.

Assembled Package Samples

Flip-chip, CSP, and BGA assemblies can be evaluated directly or after cross-sectioning to measure the in-situ expansion behavior of the cured underfill layer between die and substrate. Cross-sectioned samples reveal how filler settling, voids, and incomplete capillary flow alter local expansion behavior compared with the bulk material.

Aged and Failed Samples

Samples exposed to thermal cycling, moisture preconditioning (85 °C / 85 % RH), high-temperature storage, or reflow simulation are analyzed to quantify how degradation, hydrolysis, or post-curing shifts CTE and Tg. These comparisons are central to failure analysis when delamination or solder fatigue is observed in the field.

Testing Items

Thermal expansion analysis of underfill adhesives typically covers the following parameters, reported with test temperature ranges and heating rates:

• Coefficient of thermal expansion below Tg (α1), expressed in ppm/°C over a specified interval such as 25–100 °C
• Coefficient of thermal expansion above Tg (α2), expressed in ppm/°C over a specified interval such as Tg+20 °C to Tg+50 °C
• Glass transition temperature (Tg) determined by TMA inflection, DSC midpoint, or DMA tan delta peak
• Ratio of α2 to α1, an indicator of expansion stability across the transition
• Onset and extrapolated end temperatures of the transition region
• Storage modulus, loss modulus, and tan delta as functions of temperature where DMA is performed
• Dimensional change during cure, cure-induced shrinkage, and post-cure expansion behavior
• Moisture-conditioned CTE and Tg to evaluate plasticization effects in humid environments

Testing Methods

Thermomechanical Analysis (TMA)

TMA is the primary method for underfill CTE measurement and is performed in accordance with ASTM E831, ISO 11359-2, or IPC-TM-650 method 2.4.41. A specimen with flat, parallel faces (typically 3–6 mm wide and 0.5–3 mm thick) rests on a quartz platform while a low-load expansion probe tracks dimensional change. The furnace ramps at 5 °C/min, commonly from -50 °C to 250 °C under nitrogen or dry air. The mean CTE is calculated from the slope of the dimension-versus-temperature curve over defined intervals, while Tg is taken as the extrapolated onset of the slope change per ASTM E1545 or ISO 11359-2.

Push-Rod Dilatometry

Dynamic Mechanical Analysis (DMA)

DMA per ASTM D4065, ASTM E1640, and ISO 6721-11 characterizes the viscoelastic transition that accompanies thermal expansion. Specimens are tested in single or dual cantilever, three-point bending, or tension at 1 Hz with a small strain amplitude while temperature ramps at 2–5 °C/min. The storage modulus drop and tan delta peak identify Tg, and the modulus values above and below Tg explain why effective expansion resistance collapses once the underfill enters the rubbery state.

Differential Scanning Calorimetry (DSC)

DSC per ISO 11357-2 or ASTM E1356 determines the calorimetric Tg midpoint and detects residual cure exotherms. Incomplete cure is a frequent root cause of unexpectedly low Tg and high α2, so DSC is routinely paired with TMA during material qualification.

Thermal Cycling Verification

Dimensional stability under cyclic loading is verified using temperature cycling profiles consistent with JEDEC JESD22-A104 (for example, -40 °C to +125 °C, 500 to 1000 cycles), followed by CTE re-measurement, cross-sectional inspection, and dye-and-pry or scanning acoustic microscopy checks for underfill delamination.

Applications

Thermal expansion testing of underfill adhesives supports design, qualification, and failure analysis across the electronics industry:

• Flip-chip in package and flip-chip on board, where CTE matching directly determines solder joint fatigue life
• Wafer-level chip-scale packages and fan-out packages, including molded underfill formulations with high silica loading
• Corner and edge bonding adhesives for large BGA devices that experience board-level bending and thermal stress
• 2.5D and 3D package integration, where underfill around microbumps and through-silicon via structures must tolerate extreme thermal gradients
• Automotive and industrial electronics, where operating ranges of -40 °C to +150 °C demand high Tg and stable α1/α2 behavior
• Material incoming inspection and supplier comparison for contract assemblers and semiconductor packaging houses
• Failure analysis of delaminated fillets, solder extrusion, die cracking, and popcorn cracking after reflow or humidity exposure

FAQ

What is a typical CTE value for a qualified underfill adhesive?

Most commercial capillary underfills exhibit α1 of 25–45 ppm/°C, deliberately positioned between the CTE of silicon and that of solder or substrate. Above Tg, α2 commonly reaches 80–150 ppm/°C, which is acceptable only if Tg remains safely above the maximum operating temperature, typically at least 110–150 °C depending on the application.

Why must Tg always be evaluated together with CTE?

CTE data are meaningless without Tg context because the material behaves as two different systems on either side of the transition. A low α1 value loses all protective benefit if Tg sits at 90 °C and the device routinely operates or reflows above that point, at which time α2 dominates and stress transfer to the solder joints increases sharply.

How do moisture and cure state affect the results?

Moisture plasticizes epoxy networks and can depress Tg by 10–30 °C while raising apparent CTE, so humidity-preconditioned testing is recommended for humid-environment products. Incomplete cure similarly lowers Tg, elevates α2, and leaves a residual exotherm detectable by DSC; for this reason, full cure verification should precede any CTE acceptance decision.

What specimen quality is required for accurate TMA results?

Specimens must have flat, parallel, ground faces free of voids and unmixed filler clusters, with a thickness between 0.5 and 3 mm. Even minor face tilt or porosity introduces artifacts in the expansion curve, so careful machining and visual or X-ray inspection of specimens before testing is standard practice.

Which method should be selected for first-pass qualification?

TMA per ASTM E831 or ISO 11359-2 combined with DSC Tg per ISO 11357-2 is the standard qualification package, supplemented by DMA when modulus evolution and damping behavior across Tg are also required. Dilatometry and thermal cycling verification are added for high-reliability automotive, medical, and aerospace programs.