Determination of the Low-Temperature Coefficient of Thermal Expansion of Alumina Substrates: Testing

The low-temperature coefficient of thermal expansion (CTE) of alumina substrates is a critical parameter for electronic packaging, cryogenic devices, and aerospace applications, where thermal mismatch stresses can cause cracking, delamination, and solder joint failure. This article describes the standard test methods—push-rod dilatometry, thermomechanical analysis, and interferometric techniques—used by accredited third-party laboratories to measure the linear thermal expansion of alumina ceramic substrates from cryogenic temperatures up to room temperature. It also covers sample preparation, data reduction, typical values, and practical guidance for engineers specifying low-temperature CTE testing.

Technical Overview

Alumina (Al₂O₃) ceramic substrates are widely used as the mechanical and electrical foundation of thick-film hybrid circuits, power semiconductor modules, direct-bond-copper (DBC) assemblies, LED packages, and sensor carriers. In most of these applications, the substrate is joined to materials with very different expansion behavior, such as silicon (approximately 2.6 ppm/K), copper (approximately 16.5 ppm/K), and solder alloys. When such assemblies are exposed to low temperatures—during cold-start operation, high-altitude flight, satellite orbits, or cryogenic service—the mismatch in contraction generates stresses at every interface. An accurate knowledge of the substrate's low-temperature coefficient of thermal expansion is therefore essential for reliability prediction and design.

At room temperature and above, the mean linear CTE of 96% to 99.8% alumina is typically quoted as 6.5 to 7.8 × 10⁻⁶/K between 25 °C and 200 °C. However, the CTE of alumina, like that of all crystalline ceramics, decreases significantly as temperature falls. Near -100 °C the instantaneous CTE drops to roughly 4 to 5 × 10⁻⁶/K, and below liquid-nitrogen temperature (77 K, or -196 °C) it falls rapidly toward zero as required by the third law of thermodynamics. Designing with room-temperature CTE values alone can therefore underestimate or overestimate contraction behavior in cryogenic and sub-zero environments, which is why dedicated low-temperature expansion measurements are performed.

Low-temperature CTE determination is typically carried out by push-rod dilatometry in accordance with ASTM E228, ISO 17563, and GB/T 16535, by thermomechanical analysis (TMA) in accordance with ASTM E831 and JIS R 1618, or by high-precision interferometric methods in accordance with ASTM E289. Accredited third-party laboratories routinely offer these measurements over ranges such as -150 °C to 25 °C, -196 °C to 25 °C, or down to 4 K for specialized cryogenic programs.

Testing Samples

96% Alumina Substrates (As-Fired or Ground)

Standard-grade 96% alumina tape-cast substrates, commonly available in thicknesses of 0.25 mm, 0.38 mm, 0.5 mm, 0.635 mm, and 1.0 mm, are the most frequently tested materials. The silicate grain-boundary phase present in this grade has a modest influence on expansion behavior, so representative sampling from the actual production lot is recommended.

99% to 99.8% Alumina Substrates (Thin-Film Grade)

High-purity, polished substrates intended for thin-film metallization, RF circuits, and precision resistors exhibit slightly higher and more reproducible CTE values closer to those of dense sintered corundum. These are often tested for aerospace and cryogenic instrumentation programs where tighter data uncertainty is required.

Sample Geometry and Preparation

• For push-rod dilatometry, samples are cut into rectangular bars, typically 6 mm to 25 mm in length, with cross-sections of 4 mm to 10 mm; longer samples are preferred because they produce a larger, more resolvable length-change signal for a given CTE.
• Both end faces must be flat, parallel, and perpendicular to the measurement axis; a parallelism tolerance of better than 0.01 mm is recommended, achieved by precision grinding or lapping.
• For thin substrates where through-thickness (out-of-plane) expansion is of interest, several identical pieces may be stacked and lightly clamped, or the test may be performed in TMA expansion mode using the full substrate thickness.
• Samples should be cleaned with solvent, dried at 150 °C for 1 to 2 hours to remove adsorbed moisture, and allowed to cool in a desiccator before measurement.
• Because polycrystalline alumina is effectively isotropic in the bulk but tape-cast substrates can show slight texture, critical programs may require separate in-plane (X/Y) and out-of-plane (Z) specimens.

Testing Items

A complete low-temperature expansion characterization of an alumina substrate typically includes the following deliverables:

• Mean (secant) coefficient of linear thermal expansion, αm, reported over defined intervals such as -196 °C to 25 °C, -150 °C to 25 °C, -100 °C to 25 °C, -65 °C to 25 °C, and -40 °C to 25 °C, expressed in ppm/K or 10⁻⁶/K.
• Instantaneous (tangential) coefficient of expansion, α(T), as a continuous function of temperature, obtained by differentiation of the expansion curve.
• The raw thermal expansion curve, ΔL/L₀ versus temperature, covering both cooling and heating branches so that hysteresis and repeatability can be assessed.
• Dimensional contraction data, useful directly for mechanical fit and gap calculations in cryogenic assemblies.
• Measurement uncertainty statements, typically on the order of ±3% to ±5% on the mean CTE or ±0.05 to 0.1 ppm/K, whichever is larger, for a well-controlled dilatometric test.

Testing Methods

Push-Rod Dilatometry (ASTM E228 / ISO 17563 / GB/T 16535)

Push-rod dilatometry is the primary and most widely accepted method for determining the linear thermal expansion of alumina substrates at low temperature. The specimen is placed in a low-temperature furnace equipped with a liquid-nitrogen cooling system or mechanical cryocooler, and a spring-loaded push-rod, usually made of fused silica or a low-expansion ceramic, transmits the specimen's length change to a high-resolution displacement sensor such as a linear variable differential transformer (LVDT) with nanometer-level resolution. Key aspects of the procedure are:

• System calibration and blank correction are performed using certified reference materials such as sapphire, fused silica, or tungsten, and the inherent expansion of the sample holder and push-rod is subtracted by an empty-run (baseline) correction.
• The specimen is cooled to the lowest target temperature, typically -150 °C or -196 °C, and held until thermal equilibration is confirmed by a stable displacement and temperature reading.
• The measurement is executed at a controlled heating rate of 2 to 5 K/min while the furnace is continuously purged with dry helium or nitrogen to ensure uniform heat exchange and to prevent moisture condensation.
• Specimen temperature is measured by a thermocouple or calibrated resistance thermometer located immediately adjacent to the sample; thermocouples must be calibrated at low-temperature fixed points, such as the liquid-nitrogen point, for cryogenic work.
• A second (repeat) cycle is commonly run to verify reproducibility; ceramic substrates should show negligible hysteresis, and any drift indicates residual stress relief, moisture effects, or fixture problems.

Thermomechanical Analysis, TMA (ASTM E831 / ISO 11359-2 / JIS R 1618)

Thermomechanical analysis uses a miniature furnace and a quartz or alumina stage-and-probe assembly in which a flat-tipped probe rests on the specimen under a small static force. The method is well suited to thin alumina substrates in either the expansion configuration (force perpendicular to the substrate face, giving through-thickness CTE) or the tension/bending configuration with adapted fixtures for in-plane measurements. TMA offers small sample size, fast turnaround, and excellent low-temperature control down to approximately -150 °C with liquid-nitrogen cooling, although the absolute accuracy is generally somewhat lower than that of a well-calibrated push-rod dilatometer because of the shorter effective specimen length. TMA is frequently used for quality-control screening and lot acceptance, while push-rod dilatometry is preferred for reference data.

Interferometric Methods (ASTM E289)

Optical interferometric dilatometry measures specimen length change by counting interference fringes between mirrors attached to the specimen ends, achieving displacement resolution on the order of a fraction of the light wavelength. This method provides the highest accuracy for very low expansion values and for cryogenic measurements where the total length change of an alumina bar may be only a few hundred nanometers per millimeter of length. Absolute interferometric dilatometers operating at 77 K or 4 K are found in national metrology institutes and in specialized accredited third-party laboratories serving superconducting electronics and space-optics programs.

Capacitive and Other Cryogenic Techniques

For measurements below 77 K, such as qualification of substrates for superconducting circuits operating at 4.2 K, three-terminal guarded capacitance dilatometers and double-dilatometer differential arrangements are used. These techniques cancel cell expansion by measuring the specimen against a reference of known expansion and can resolve CTE values below 0.1 ppm/K, where mechanical push-rod systems lose sensitivity.

Data Reduction

The mean coefficient of linear thermal expansion between two temperatures T₁ and T₂ is calculated from the measured lengths using the relationship αm = (L₂ − L₁) / [L₀ × (T₂ − T₁)], where L₀ is the specimen length at the reference temperature, normally 25 °C. The instantaneous coefficient at any temperature is obtained as α(T) = (1/L₀) × dL/dT, computed by numerical differentiation with appropriate smoothing. Reported results should always state the reference temperature, the temperature interval, the heating rate, the purge gas, the calibration standard used, and the measurement uncertainty.

Applications

Electronic Packaging and Power Modules

DBC and active-metal-brazed (AMB) alumina substrates in IGBT and SiC power modules experience deep thermal cycling, including cold starts at -40 °C and below. Low-temperature CTE data feed into finite-element models that predict copper-layer and solder-joint stress, guiding substrate selection and bonding design.

Aerospace, Satellite, and Defense Electronics

Equipment qualified to low-temperature limits of -55 °C or -65 °C, as well as payloads exposed to orbital thermal cycling, requires verified contraction behavior of all ceramic components to maintain alignment, hermeticity, and joint integrity.

Cryogenic and Superconducting Devices

Substrates for high-temperature superconducting (HTS) thin films, cryogenic detectors, and quantum electronics operate at 77 K or 4 K. CTE matching between the substrate, the deposited film, and mounting structures over the full cooldown range is critical to prevent film cracking and delamination.

Precision Optics and Metrology Structures

Alumina is used in instrument carriers and optical benches where dimensional stability across temperature excursions must be quantified; low-temperature expansion curves support compensation calculations in precision mechanical design.

FAQ

What temperature range is considered "low temperature" for CTE testing of alumina substrates?

In practice, "low temperature" covers any range below room temperature. Common industrial intervals are -40 °C, -55 °C/-65 °C, and -150 °C to 25 °C, while cryogenic programs specify measurement down to 77 K (-196 °C) or 4 K. The test plan should match the actual service environment of the product.

Why does the CTE of alumina decrease at low temperatures?

Thermal expansion in ceramics arises from lattice vibrations (phonons). As temperature decreases, phonon populations diminish and the anharmonicity of the interatomic potential contributes less to expansion, so both the Grüneisen parameter and the CTE fall. In accordance with the third law of thermodynamics, the CTE approaches zero as absolute zero is approached.

What are typical low-temperature CTE values for 96% and 99.6% alumina substrates?

As indicative figures, the mean CTE from 25 °C to -100 °C is approximately 5 to 5.5 ppm/K for 96% alumina and approximately 5.3 to 5.8 ppm/K for high-purity grades, compared with 6.5 to 7.5 ppm/K between room temperature and 200 °C. Values vary with composition, texture, and test method, so project-specific measurement is recommended for critical designs.

What sample size is needed, and is the test destructive?

Push-rod dilatometry generally requires a bar of 6 mm to 25 mm in length; TMA can work with pieces a few millimeters across. The measurement itself does not alter the material, but specimens must be cut from the substrate, so sample preparation is considered destructive to the panel. Laboratories typically request three to five replicate specimens per condition.

What is the difference between mean CTE and instantaneous CTE?

Mean (secant) CTE is a single averaged value over a stated temperature interval and is what most drawings and datasheets specify. Instantaneous (tangent) CTE is the point-by-point slope of the expansion curve and is needed for detailed stress analysis near cryogenic temperatures where the expansion behavior changes rapidly.

How are measurement errors controlled at low temperatures?

Error control relies on calibrated temperature sensors positioned at the specimen, blank-run correction of the dilatometer fixture, verification against certified reference materials such as sapphire, dry helium purge to eliminate temperature gradients and condensation, slow controlled heating rates, and duplicate thermal cycles to demonstrate repeatability.