Underfill Adhesive Heat Resistance Testing Methods: Standards, Procedures and Evaluation Criteria

Underfill adhesives play a critical role in reinforcing solder interconnects in flip-chip, BGA and CSP assemblies, and their long-term reliability is governed by thermal resistance. This technical guide explains the sample types, test items and internationally recognized testing methods used by accredited third-party laboratories to qualify the heat resistance of underfill materials. Topics covered include glass transition temperature, coefficient of thermal expansion, thermal decomposition, temperature cycling, high-temperature aging and post-exposure bond strength evaluation.

Technical Overview

Underfill adhesives are epoxy-based or epoxy-modified polymer materials dispensed between a semiconductor device and its substrate to redistribute thermo-mechanical stress away from solder joints. During solder reflow, high-temperature operation, and repeated power cycling, the underfill layer must maintain adhesion, dimensional stability, and mechanical integrity. Poor heat resistance leads to characteristic failure modes such as delamination at the die-to-underfill or underfill-to-substrate interface, solder joint cracking caused by stress relaxation, voiding from moisture expansion, and embrittlement from excessive post-cure.

Heat resistance evaluation therefore combines intrinsic material property measurements (thermal transition, expansion and decomposition behavior) with endurance testing that simulates the full product life cycle. Accredited third-party laboratories typically structure the qualification program around ISO, ASTM, IEC, GB, UL and JIS standards so that results are comparable across suppliers and accepted by downstream OEMs.

• Intrinsic properties: glass transition temperature (Tg), coefficient of thermal expansion (CTE), decomposition temperature (Td), storage modulus retention
• Endurance properties: high-temperature storage, temperature cycling, thermal shock, reflow heat exposure
• Functional retention: die shear and lap shear strength after thermal aging, delamination inspection, mass loss

Testing Samples

Bulk Cured Specimens

Liquid underfill is cast into films, plaques or pans and cured using the supplier-recommended schedule (for example, 150 °C for 30 to 60 minutes, followed by post-cure if specified). Specimens are conditioned at 23 °C and 50% relative humidity in accordance with ISO 291 before testing. Bulk samples are used for DSC, TGA, TMA and DMA measurements, typically requiring only 5 to 50 mg of material per run.

Bonded Joint Specimens

To evaluate functional heat resistance, the underfill is bonded between standardized substrates: copper-to-copper or FR-4 lap shear coupons prepared per ASTM D1002 or ISO 4587, and silicon die attached to FR-4, BT or laminate substrates for die shear testing. Bondline thickness is controlled with glass beads or spacers, commonly in the 25 to 100 µm range, to reflect real underfill gap heights.

Component-Level Assemblies

Actual flip-chip, BGA, CSP or wafer-level packages underfilled with the candidate material are used for endurance testing. These assemblies allow evaluation of the complete material stack, including flux residue interaction, flow-induced voids, and filler particle sedimentation effects that bulk specimens cannot reveal. Baseline C-SAM (acoustic microscopy) and X-ray inspection are performed before exposure to establish a defect-free reference.

Testing Items

A complete heat resistance test matrix for underfill adhesives generally includes the following items, with typical target values for electronics-grade capillary or no-flow underfills:

• Glass transition temperature (Tg): target 90–150 °C depending on application temperature class
• Coefficient of thermal expansion below Tg (alpha-1): target 25–45 ppm/°C; above Tg (alpha-2): typically below 150 ppm/°C
• Thermal decomposition temperature (Td, 5% mass loss): target above 300 °C
• Total mass loss and volatile condensables (outgassing): TML below 1%, CVCM below 0.1%
• Storage modulus at room temperature and at maximum operating temperature, and modulus retention after aging
• Heat deflection temperature (HDT) for filled systems where applicable
• High-temperature storage endurance: 125–175 °C for 500–1000 hours
• Temperature cycling endurance: -40 °C to +125 °C or -55 °C to +125 °C, 500–1000 cycles
• Thermal shock resistance using two-chamber or liquid-to-liquid transition profiles
• Reflow heat resistance: survival of one to three reflow passes at 260 °C peak without cracking or delamination
• Die shear and lap shear strength retention after thermal exposure: commonly 80% minimum retention
• Flammability rating per UL 94 (typically V-0 or V-1 at thin sections)

Testing Methods

Glass Transition Temperature by DSC

Differential scanning calorimetry is performed per ISO 11357-2, ASTM E1356, GB/T 19466.2 or JIS K 7121. A cured specimen of approximately 5–10 mg is heated under nitrogen at 10 °C/min, and Tg is reported as the midpoint of the step transition (half-height method). A second heating scan removes thermal history effects. Delta-Tg measured before and after post-cure or thermal aging also serves as a cure-completeness and degradation indicator; a downward shift greater than 10 °C after aging suggests matrix degradation.

Coefficient of Thermal Expansion by TMA

Thermomechanical analysis per ISO 11359-2 or ASTM E831 measures linear expansion through Tg at a heating rate of 5 °C/min with a low-force expansion probe. Alpha-1 is reported from 25 °C to Tg minus 20 °C, and alpha-2 above Tg. Because CTE mismatch between silicon (2.6 ppm/°C), the underfill, and the substrate drives solder joint stress, low alpha-1 is a primary selection criterion, especially for large-die flip-chip packages.

Thermal Decomposition by TGA

Thermogravimetric analysis per ISO 11358-1 or ASTM E1131 heats the cured specimen at 10 °C/min under nitrogen (with an optional air switch to quantify oxidative stability and filler content from residue). The 5% weight-loss temperature defines Td; well-formulated underfills exceed 300 °C, comfortably above the 260 °C reflow peak. Outgassing behavior relevant to voiding and contamination-sensitive optics is screened separately per ASTM E595.

Modulus and Tg Retention by DMA

Dynamic mechanical analysis per ISO 6721-1/-10 or ASTM D4065 records storage modulus and tan delta from -60 °C to 250 °C. DMA-derived Tg usually appears 10–20 °C higher than DSC values. For heat resistance, the key output is storage modulus at the maximum operating temperature (for example, 150 °C): if the modulus drops below approximately 100 MPa in the rubbery region, the underfill can no longer effectively constrain solder joints at high temperature.

Temperature Cycling and Thermal Shock

Cycling is conducted per IEC 60068-2-14 (test Nb) or the equivalent GB/T 2423.22 and JIS C 60068-2-14 procedures, with dwell times of 10–15 minutes at each extreme and ramp rates of 10–15 °C/min. Common profiles are -40 °C to +125 °C for consumer and industrial electronics and -55 °C to +125 °C for automotive grades, with 500 to 1000 cycles as standard qualification length. Faster liquid-to-liquid thermal shock (IEC 60068-2-14 test Na) accelerates screening. Endpoints include electrical continuity, C-SAM delamination mapping, dye-and-pry solder joint inspection, and cross-sectional microscopy; typical acceptance requires no interfacial delamination exceeding 10% of the bonded area and no solder cracks.

High-Temperature Storage (Dry Heat Aging)

Specimens are exposed per IEC 60068-2-2 (test Bb), GB/T 2423.2 or JIS C 60068-2-2 at 125 °C, 150 °C or 175 °C for 500 and 1000 hours. After exposure, Tg, mass loss (target below 1%), and bond strength are re-measured. For long-term thermal endurance ratings, UL 746B provides the framework for establishing the Relative Temperature Index (RTI) through multi-temperature Arrhenius-type aging of dielectric and mechanical properties.

Reflow Heat Resistance

Preconditioned assemblies are passed through reflow profiles with a peak of 260 °C per IEC 60068-2-58 soldering-heat test procedures. Materials must survive multiple passes without popcorning, cracking or delamination, since double-sided assembly and rework expose cured underfill to repeated excursions above Tg. Post-reflow inspection combines optical, acoustic and cross-sectional methods.

Bond Strength After Thermal Exposure

Lap shear strength is measured per ASTM D1002 or ISO 4587, and die shear strength using standard die shear equipment with a 100–500 kgf load capability, on both aged and unaged specimens. Shear strength retention after 1000 hours at 150 °C, or after temperature cycling, is compared against baseline; retention of 80% or greater with cohesive (rather than adhesive) failure modes is the commonly applied pass criterion.

Applications

Heat resistance test data for underfill adhesives directly support material selection and qualification in the following areas:

• Flip-chip and micro-bump packaging for high-performance computing and AI processors
• Ball grid array (BGA) and chip-scale package (CSP) corner and edge bonding
• Wafer-level chip-scale packages (WLCSP) requiring reflow-compatible underfill films
• Automotive electronics such as ADAS domain controllers, ECUs and powertrain modules operating across -40 °C to +125 °C
• 5G and millimeter-wave modules with high power density and elevated board temperatures
• Stacked-die and package-on-package (PoP) memory assemblies
• Power module encapsulation where sustained high junction temperatures demand high-Tg, low-CTE matrices

FAQ

What Tg value should an underfill adhesive have?

For most consumer electronics, a Tg of 90–130 °C measured by DSC is sufficient, provided it exceeds the maximum operating temperature. Automotive and power applications typically require 130–150 °C or higher, along with high modulus retention above Tg so the underfill continues to reinforce joints at elevated temperature.

How many samples are needed for a heat resistance evaluation?

DSC, TGA and TMA each require only milligram-scale specimens, typically three replicates per condition. Lap shear testing uses 10–20 coupons per aging condition, and die shear uses 10–25 die per leg. Endurance legs (cycling, storage) require 30–50 components per readout point to achieve statistical confidence.

How long does a 1000-cycle temperature cycling test take?

At a typical -40 °C to +125 °C profile with 10-minute dwells, one cycle lasts roughly 40–50 minutes, so 1000 cycles take approximately four to five weeks, plus inspection and mechanical readout time.

Why do DSC and TMA sometimes report different Tg values for the same underfill?

DSC detects the heat-capacity step transition, while TMA detects the expansion-rate change and DMA detects mechanical relaxation. Values can differ by 10–20 °C. Reports should always state the measurement method; for joint-reinforcement purposes, the DMA or TMA value is often the more functionally relevant indicator.

What criteria define failure in thermal endurance testing?

Commonly applied criteria include interfacial delamination visible by C-SAM exceeding 10% of the bonded area, solder joint cracks confirmed by cross-section or dye-and-pry, electrical opens or resistance shift beyond specification, mass loss above 1% after aging, and shear strength retention below 80% of baseline.

Which standard combination is recommended for a full qualification report?

A typical accredited third-party laboratory report combines ISO 11357-2/ASTM E1356 (Tg), ISO 11359-2/ASTM E831 (CTE), ISO 11358/ASTM E1131 (decomposition), ISO 6721/ASTM D4065 (DMA), IEC 60068-2-2 (dry heat), IEC 60068-2-14 (temperature cycling and shock), IEC 60068-2-58 (reflow resistance), ASTM D1002/ISO 4587 (lap shear) and UL 94 (flammability), giving OEMs a complete and comparable thermal performance picture.