Determination of Martens Heat Resistance Temperature of Underfill Adhesives

The Martens heat resistance temperature is a key thermal-mechanical index that evaluates how a cured underfill adhesive behaves under sustained bending stress at elevated temperature. This article explains the testing principle, specimen preparation, apparatus, procedure, and data interpretation for determining the Martens temperature of underfill adhesives in accordance with recognized standards such as GB/T 1699. It is intended for semiconductor packaging engineers, materials scientists, and quality teams who need to select, qualify, and monitor underfill materials with confidence.

Technical Overview

Underfill adhesives are low-viscosity thermosetting materials—most commonly epoxy resins filled with high loadings of spherical silica—that are introduced beneath flip-chip, ball grid array (BGA), and chip-scale package (CSP) devices to protect solder interconnects. After curing, the underfill forms a rigid composite that redistributes thermomechanical stress between the silicon die and the organic substrate, suppressing solder fatigue and extending package service life. Because these assemblies routinely experience reflow, burn-in, and sustained high-temperature operation, the heat resistance of the cured underfill under mechanical load is a first-order reliability parameter.

The Martens heat resistance temperature, often simply called the Martens temperature, is the temperature at which a standard cured bar specimen, mounted in a Martens apparatus and loaded to a defined bending stress, reaches a specified deflection while being heated at a constant rate. The method, specified for rigid plastics in GB/T 1699, identifies the point at which thermal softening causes marked deformation under load. For cured underfill formulations, the Martens temperature is closely associated with the glass transition region and provides a practical, comparative measure of heat resistance under load.

Unlike the glass transition temperature (Tg) measured by DSC or DMA, which is an intrinsic material property, the Martens temperature is an engineering index that combines modulus decay, thermal softening, and creep behavior under a fixed stress. It is therefore widely used for formulation screening, supplier datasheet verification, incoming material control, and failure analysis. Accredited third-party laboratories routinely perform this determination alongside complementary measurements such as heat deflection temperature (ISO 75 / GB/T 1634), Vicat softening temperature (ISO 306 / GB/T 8802), and Tg by DSC (GB/T 19466 / ISO 11357).

Testing Samples

Capillary Underfill Adhesives

Liquid epoxy systems designed to flow by capillary action into the gap between die and substrate are cast into bar molds and cured using the supplier's recommended thermal profile, typically one to three hours at 100–165 °C with a post-cure where specified. Casting and curing the bulk liquid under production-equivalent conditions ensures the tested bar represents the material state achieved in actual packages.

No-Flow Underfill Adhesives

Fluxing epoxy-based pastes or films, which cure during the solder reflow step, are cast into molds and passed through the same reflow profile used in assembly. Because no-flow systems cure simultaneously with solder melting, matching the actual reflow profile is essential for obtaining a representative degree of cure.

Molded and Wafer-Level Underfill Materials

Silica-filled molding compounds used as molded underfill, as well as wafer-applied coatings, are molded, compression-formed, or coated into flat plaques and fully cured. Test bars are then cut or machined from the cured plaques, taking care to avoid edge chips, cracks, or overheating during machining.

Specimen Requirements

The standard Martens specimen is a rectangular bar of approximately 120 mm × 15 mm × 10 mm, flat, stress-free, and free of voids, bubbles, sink marks, and edge defects. Before testing, specimens are conditioned in a standard laboratory atmosphere of 23 ± 2 °C and 50 ± 5 % relative humidity (ISO 291 / GB/T 2918) for a defined period, typically not less than 24 hours, unless otherwise agreed. A minimum of three valid specimens per material is recommended so that mean values and scatter can be assessed.

Testing Items

The primary deliverable of the determination is the Martens heat resistance temperature of the cured underfill adhesive, expressed in degrees Celsius. Supporting data are normally supplied with the result to ensure traceability and comparability.

• Martens heat resistance temperature (°C) for each specimen
• Deflection-versus-temperature curve or data points for each specimen
• Mean value, range, and standard deviation across valid specimens
• Cure schedule, post-cure condition, and conditioning history applied to the bars
• Any anomalies such as cracking, warping, or filler sedimentation observed during the test

Where a fuller thermal-mechanical picture is required, the determination is frequently bundled with heat deflection temperature (ISO 75-2), Vicat softening temperature (ISO 306), Tg by DSC or DMA, and coefficient of thermal expansion measurements to characterize the underfill completely.

Testing Methods

Apparatus

The test is performed on a Martens heat resistance tester comprising a rigid specimen clamp that holds the bar as a loaded cantilever, a lever-and-weight system calibrated to impose a nominal bending stress of approximately 5 MPa, and a displacement indicator graduated to 0.01 mm for reading deflection. The loaded assembly is immersed in a stirred, heat-transfer liquid bath (typically silicone oil) equipped with programmed heating capable of maintaining a constant rate of 50 °C per hour, and a calibrated thermocouple or precision thermometer positioned immediately adjacent to the critical section of the specimen.

Specimen Preparation

Liquid underfill is de-aerated if necessary, cast into pre-heated molds, and cured per the manufacturer's schedule, including any specified post-cure. After demolding, bars are inspected dimensionally and visually, conditioned per ISO 291 / GB/T 2918, and marked for identification. The curing details are recorded, because residual under-cure or moisture uptake in epoxy systems can measurably depress the Martens temperature.

Test Procedure

• Measure and record the specimen dimensions at three points along its length and verify freedom from defects.
• Mount the specimen vertically in the Martens apparatus, attach the loading lever, and apply weights calibrated to produce the specified 5 MPa bending stress at the surface fibers.
• Set the deflection indicator to zero or record the initial reading, then immerse the assembly in the bath at a starting temperature comfortably below the expected result, commonly room temperature.
• Heat the bath uniformly at 50 ± 5 °C per hour with continuous stirring, ensuring the temperature gradient along the specimen is minimized.
• Record deflection and temperature at regular intervals; the Martens heat resistance temperature is the bath temperature at which the deflection indicator registers the specified endpoint of 6 mm.
• Repeat for the remaining specimens; discard and replace any specimen that cracks, slips in the clamp, or softens without a clear deflection endpoint.

Data Processing and Reporting

Individual Martens temperatures are reported for each valid specimen together with the arithmetic mean; if replicate values differ beyond the agreed tolerance, additional specimens are tested to establish a reliable mean. The test report should identify the material and lot, cure and post-cure conditions, specimen dimensions, applied stress, heating rate, endpoint deflection, conditioning history, the standard applied, and the apparatus used. This completeness allows direct comparison of results obtained at different times or by different accredited third-party laboratories.

Applications

Martens heat resistance temperature data for underfill adhesives are used across the electronics packaging value chain, from resin formulation to field failure analysis. The measurement supports material selection, ensures that cured adhesives will survive subsequent high-temperature processes, and provides objective evidence for datasheet compliance.

• Qualification of capillary, no-flow, and molded underfills for flip-chip, BGA, CSP, and wafer-level packages
• Verification of incoming material lots against supplier datasheet specifications
• Formulation development and benchmarking of epoxy hardness, filler loading, and crosslink density
• Confirmation that the underfill withstands downstream processes such as molding, rework exposure, burn-in, and high-temperature operation
• Support for automotive, industrial, and high-reliability programs where sustained elevated-temperature service is specified
• Failure analysis root-cause studies where thermal softening or insufficient cure of the underfill is suspected

FAQ

How does the Martens temperature differ from Tg, HDT, and Vicat softening temperature?

All four characterize thermal softening but are not interchangeable. Tg from DSC or DMA is an intrinsic transition temperature measured without external load. HDT (ISO 75 / GB/T 1634) applies bending stresses of 1.80 MPa, 0.45 MPa, or 8.00 MPa to a simply supported bar and uses small deflection endpoints. The Martens test applies approximately 5 MPa bending stress to a cantilevered bar and uses a 6 mm deflection endpoint. Vicat softening (ISO 306) uses a needle penetrator under 10 N or 50 N. Results generally trend together for related formulations, but values must never be compared across methods.

The underfill is supplied as a liquid—how are solid test bars produced?

The liquid adhesive is cast into bar-shaped molds of the required dimensions and cured using the same thermal profile specified for production assemblies, including any post-cure. Because the Martens temperature depends strongly on the degree of cure, replicating the real cure schedule is essential for a meaningful result.

How much material is needed for testing?

In practice, roughly 30–50 g of liquid underfill is sufficient to cast a set of three to five standard bars. For molded or film-form underfills, a cured plaque large enough to yield at least three bars is required.

What heating rate and endpoint are used?

The bath is heated at a constant 50 °C per hour, and the Martens temperature is recorded as the temperature at which the deflection indicator reaches 6 mm under the applied 5 MPa bending stress, in accordance with GB/T 1699.

What Martens temperature should an underfill adhesive have?

It depends on the application. As a guideline, the Martens temperature should comfortably exceed the highest sustained operating temperature of the package plus a safety margin. Well-cured, highly filled epoxy underfills commonly exhibit Martens temperatures in the approximate range of 130–200 °C, consistent with typical datasheet Tg values; automotive and high-temperature applications demand the upper end of this range.

Does a passing Martens result replace package-level reliability testing?

No. The Martens temperature is a bulk material index that supports material screening and acceptance. Full reliability assurance still requires package-level assessments such as temperature cycling, thermal shock, and humidity aging, in which the underfill is tested in its actual geometry and stress state.