IC Substrate Tg and Thermal Expansion Measurement: Glass Transition Temperature and CTE Testing by T

IC substrate Tg point expansion measurement determines the glass transition temperature (Tg) and coefficient of thermal expansion (CTE) of package substrates through thermomechanical analysis (TMA). This testing is critical for evaluating dimensional stability, plated via integrity, and solder joint reliability during reflow soldering and thermal cycling. This article explains the test principle, sample requirements, methodology in accordance with IPC-TM-650, typical specification limits, and practical applications for IC substrate qualification.

Technical Overview

The glass transition temperature (Tg) is the temperature at which an organic dielectric material in an IC substrate changes from a rigid, glassy state to a softer, more rubbery state. Above Tg, the resin matrix undergoes a significant increase in molecular mobility, which causes an abrupt acceleration in thermal expansion. Thermomechanical analysis (TMA) detects this transition by continuously monitoring the thickness of a specimen while it is heated at a controlled rate under a minimal probe load.

For IC package substrates, expansion in the Z-axis (through-thickness) direction is a dominant reliability concern. As the substrate expands and contracts across the Tg, copper-plated vias experience cyclic tensile and shear stress, which can initiate cracking, interfacial delamination, and insulation failure. Accurate Tg and CTE data are therefore essential inputs for material selection, process qualification, and warpage simulation.

• Tg is determined from the TMA expansion curve as the intersection of the extrapolated tangents drawn below and above the transition region.
• Alpha 1 (CTE below Tg) is calculated from the linear slope of the expansion curve below the transition, expressed in ppm/degree Celsius.
• Alpha 2 (CTE above Tg) is calculated from the slope above the transition and is typically several times larger than alpha 1.
• Total thickness expansion, expressed as a percentage over a defined range such as 25 to 260 degrees Celsius, reflects behavior during lead-free reflow.

Testing Samples

Bare Core Laminates

Core materials such as bismaleimide triazine (BT) resin, epoxy-glass, and modified epoxy systems are tested in laminate form, typically with thickness between 0.10 and 0.80 mm. Specimens are cut into squares or rectangles of approximately 6 to 10 mm per side to ensure a stable, flat contact surface for the quartz expansion probe.

Build-up Film Materials

Photo-imageable dielectric and build-up films, such as Ajinomoto build-up film (ABF) and similar resin systems, are tested as laminated stacks on carrier cores or as free-standing films. Because individual film layers are thin (15 to 100 micrometers), multiple layers are often stacked to reach a measurable total thickness that provides acceptable resolution.

Finished Multilayer Substrates

Completed package substrates, including through-hole, high-density interconnect, and embedded via constructions, can be measured to characterize the composite Z-axis behavior of the full stack. Sampling locations are selected in resin-rich areas or regions free of dense circuitry to minimize local distortion from copper distribution.

All specimens should be baked (typically 105 to 120 degrees Celsius for 2 hours) to remove absorbed moisture and stored in a desiccator prior to testing, since moisture can depress the measured Tg and inflate the apparent expansion.

Testing Items

A standard Tg point expansion test report for IC substrates includes the following measured parameters:

• Tg by TMA, reported as the tangent-intersection temperature in degrees Celsius.
• Alpha 1 CTE (below Tg), in ppm/degree Celsius, over a specified reference range such as 50 to 150 degrees Celsius or 30 degrees Celsius to Tg minus 20 degrees Celsius.
• Alpha 2 CTE (above Tg), in ppm/degree Celsius, from Tg plus 20 degrees Celsius to the upper test limit.
• Total thickness expansion in percent, commonly reported from 25 to 260 degrees Celsius for reflow compatibility assessment.
• Optional second-scan Tg to evaluate cure completeness and thermal history effects.
• Optional comparative results after moisture sensitivity level preconditioning and reflow simulation to assess delamination resistance.

Testing Methods

TMA Expansion Probe Method

The primary method follows IPC-TM-650 method 2.4.24 for Tg and method 2.4.41 for Z-axis CTE. A polished specimen is placed on the quartz stage of the thermomechanical analyzer, and a flat-tipped expansion probe applies a small static force, typically on the order of a few grams, to maintain contact without deforming the sample. The furnace is purged with nitrogen or dry helium and heated at a controlled rate, typically 5 degrees Celsius per minute, from room temperature to a maximum of 250 to 300 degrees Celsius, ensuring the material is driven well above Tg while remaining below its decomposition temperature.

Data Interpretation

The instrument records dimensional change as a function of temperature. The CTE in each region is derived from the slope of the linear segments using the relationship of dimensional change divided by original thickness and temperature interval. Tg is taken at the intersection of the extrapolated pre-transition and post-transition tangent lines. When evaluating cure state or removing thermal history, a first heating scan is used for conditioning and the second scan is reported, following common industry practice for resin systems.

Complementary Techniques

Differential scanning calorimetry (IPC-TM-650 method 2.4.25) detects Tg from the step change in heat capacity and is well suited to resin-rich specimens, while dynamic mechanical analysis (method 2.4.24.2) detects Tg from the peak in mechanical damping and typically returns the highest reported Tg value. Because each technique responds to a different physical manifestation of the transition, results differ systematically, and the measurement method must always be stated alongside the reported value.

Applications

Tg point expansion measurement supports the entire IC substrate development and manufacturing lifecycle:

• Incoming material qualification of core laminates and build-up films against supplier specifications and procurement standards.
• Lamination and post-cure process optimization, where insufficient cure is revealed by depressed Tg and elevated alpha 2 values.
• Reliability engineering for plated via integrity, via barrel cracking, and pad lifting during thermal cycling.
• Solder joint fatigue and reflow compatibility assessment for ball grid array and flip-chip package assemblies.
• Warpage prediction and finite element simulation, where accurate Z-axis CTE data are mandatory inputs.
• Failure analysis to confirm whether out-of-spec thermal expansion contributed to field or assembly defects.
• Lot-to-lot statistical process control and material supplier comparison programs.

FAQ

Why do TMA, DSC, and DMA give different Tg values for the same substrate?

The three techniques measure different physical responses to the glass transition. DMA detects mechanical relaxation and typically reports the highest Tg, DSC detects the heat capacity step change and reports an intermediate value, and TMA detects the change in expansion slope and generally reports the lowest value. Differences of 10 to 30 degrees Celsius among methods are normal, so the test method must always accompany the reported number.

Why is Z-axis expansion more critical than in-plane expansion for IC substrates?

Plated through-holes and via structures span the thickness of the substrate. When Z-axis expansion accelerates above Tg, the copper barrels are stretched disproportionately relative to the surrounding laminate, concentrating stress at via corners and interfaces. In-plane expansion, constrained largely by glass fabric and copper, is usually much lower and is primarily a warpage concern rather than a via reliability concern.

What are typical specification values for IC substrate materials?

High-performance BT cores commonly exhibit Tg in the range of 180 to 220 degrees Celsius with alpha 1 around 15 to 25 ppm/degree Celsius, while build-up films typically show alpha 1 between 20 and 45 ppm/degree Celsius with alpha 2 of 100 ppm/degree Celsius or higher. Exact limits are set by each product specification, and total expansion from 25 to 260 degrees Celsius is often restricted to a few percent for fine-line applications.

What sample preparation precautions affect accuracy?

Specimens must be flat, parallel, and free of burrs, with adequate thickness for resolution, typically 0.5 mm or more for laminate testing or stacked layers for thin films. Moisture must be removed by pre-baking, the probe must rest on a uniform resin area away from dense copper features, and a consistent heating rate must be maintained so that results are comparable between runs and laboratories.