Thermomechanical analysis (TMA) of alumina substrates is a critical characterization technique for measuring the coefficient of thermal expansion (CTE), dimensional stability, and high-temperature mechanical behavior of Al2O3 ceramic substrates used in electronics packaging. This article provides a comprehensive technical overview of alumina substrate TMA testing, covering sample preparation, testing items, experimental methods in accordance with ASTM E831, ISO 11359, and ISO 17561 standards, typical applications, and frequently asked questions for engineers and quality assurance professionals.
Technical Overview
Alumina (Al2O3) ceramic substrates are among the most widely used dielectric materials in electronic packaging, thick-film and thin-film hybrid circuits, power semiconductor modules, and LED assemblies. Because these substrates are routinely bonded to copper, silicon, solder, and other materials with dissimilar thermal expansion behavior, the coefficient of thermal expansion (CTE) of the alumina must be precisely known to predict thermo-mechanical stress, delamination risk, and long-term reliability of the assembled device.
Thermomechanical analysis (TMA) is a thermal analysis technique in which the dimensional change of a specimen is monitored while it is subjected to a controlled temperature program under a small static mechanical load. For alumina substrates, TMA experiments deliver quantitative data on linear thermal expansion, CTE over specified temperature intervals, sintering or densification shrinkage of green tapes, softening behavior of glassy grain-boundary phases, and anisotropic expansion between the in-plane (X/Y) and thickness (Z) directions. Typical CTE values measured for high-purity substrates range from approximately 6.5 to 8.0 ppm/°C between room temperature and 800 °C, depending on alumina content (96% vs. 99.5% vs. 99.6%), crystallographic texture, and glass-phase fraction.
Testing is commonly performed by accredited third-party laboratories and in-house reliability labs following recognized standards such as ASTM E831 (linear thermal expansion of solid materials by TMA), ASTM E228 (push-rod dilatometry), ISO 11359-2 (thermomechanical analysis), and ISO 17561 (linear thermal expansion of fine ceramics). These standards define instrument calibration, dimensional requirements, heating rates, and data-reduction procedures to ensure repeatable, interlaboratory-comparable results.
Testing Samples
As-Fired 96% Alumina Substrates
Standard 96% alumina substrates used for thick-film hybrid circuits and DBC (direct bonded copper) power modules. The residual glassy silicate phase makes these samples sensitive to softening effects above approximately 900 °C, which the TMA experiment can detect as an abrupt change in expansion slope. Specimens are typically cut to 5–10 mm squares or strips from production lots for lot-release testing.
High-Purity 99.5%–99.6% Alumina Substrates
Substrates intended for thin-film metallization, RF circuits, and precision resistor networks. These samples exhibit highly linear, reproducible expansion curves and serve as reference materials for validating instrument performance. CTE anisotropy between the pressing direction and the plane of the tape is frequently characterized for these grades.
Green Tape and Unfired Ceramic Sheets
Tape-cast alumina sheets containing organic binders and plasticizers are tested in dedicated TMA runs to measure binder burn-out shrinkage, sintering onset temperature, and total firing shrinkage. This information is essential for designing co-fired metallization layouts in HTCC (high-temperature co-fired ceramic) packages.
Metallized and Coated Substrates
Substrates carrying screen-printed conductors, tungsten–molybdenum metallization, or glass passivation layers can be evaluated to determine the composite expansion behavior and to identify stress-induced warpage or interfacial delamination during thermal cycling within the TMA furnace.
Sample Preparation Requirements
- Faces of the specimen must be flat and parallel; surface grinding or lapping may be required when cut edges create non-parallel loading surfaces.
- Recommended specimen height (dimension along the measurement axis) is 2–10 mm; typical substrate chips of 4–6 mm stacked thickness are acceptable.
- Specimens must be dried at 110–150 °C for 1–2 hours and stored in a desiccator to remove adsorbed moisture that would otherwise distort the baseline.
- Chipped, warped, or cracked samples must be excluded, as defects cause probe seating errors and erratic displacement signals.
- At least three specimens per material lot and per orientation (in-plane and through-thickness) are recommended for statistical validity.
Testing Items
- Mean coefficient of linear thermal expansion (CTE) over defined intervals, e.g., 25–200 °C, 25–300 °C, 25–500 °C, and 25–800 °C.
- Instantaneous (differential) CTE as a continuous function of temperature.
- Total linear expansion (ΔL/L0, %) across the programmed temperature range.
- Sintering onset temperature, maximum shrinkage rate temperature, and total firing shrinkage for green tapes.
- Softening point or glass-transition-like inflection of grain-boundary glass phases in 96% alumina.
- CTE anisotropy ratio between the in-plane (X/Y) and through-thickness (Z) directions.
- Dimensional stability and irreversible length change after thermal exposure (residual shrinkage on second heat run).
- Creep or penetration behavior under elevated static load for substrates with significant glass content.
Testing Methods
Instrumentation and Setup
A thermomechanical analyzer fitted with a high-temperature furnace (typically up to 1000–1100 °C for alumina work), a fused silica or high-purity alumina sample stage and probe, and a linear variable displacement transducer (LVDT) with resolution of ±0.1 µm or better is used. The experiment is conducted under a flowing high-purity nitrogen or dry air atmosphere (20–50 mL/min) to prevent probe oxidation and convective signal noise. A small static force of 0.02–0.2 N is applied via the expansion (flat) probe to maintain contact without deforming the specimen.
Calibration
Per ASTM E831 and ISO 11359-2, the instrument must be calibrated immediately before the test campaign: temperature calibration using the Curie point of nickel or the melting point of indium and zinc standards; length-change calibration using a certified reference material such as sapphire (corundum) or fused silica whose expansion curve is traceable to a national metrology institute. Baseline correction is performed by running an empty-stage or standard-material dummy run to subtract system expansion.
Experimental Procedure
- Measure the initial specimen length (L0) at 23 ± 2 °C with a calibrated micrometer to ±2 µm, or rely on instrument entry values as permitted by the standard.
- Position the specimen centered under the probe, apply the specified static force, and allow the displacement signal to stabilize.
- Equilibrate at the starting temperature (typically 20–30 °C) for 5–10 minutes, then heat at a constant rate of 3–5 °C/min to the maximum test temperature (commonly 800–1000 °C for metallized substrates).
- Optionally record the cooling segment to evaluate hysteresis and irreversible dimensional change.
- Calculate the mean CTE using α = ΔL / (L0 × ΔT), where ΔL is the length change over the temperature interval ΔT, and report values in ppm/°C (×10⁻⁶ /°C) to three significant figures.
Ancillary Techniques
Where TMA sensitivity is insufficient for very thin substrates, push-rod dilatometry per ASTM E228 or ISO 17561 may be employed, offering excellent accuracy for CTE above 5 ppm/°C. Complementary methods such as thermogravimetric analysis (TGA) for binder burn-out, dilatometry of cofired test coupons, and cyclic TMA runs for reliability screening are frequently integrated into a full characterization plan by accredited third-party laboratories.
Applications
- CTE matching between alumina substrates and silicon dies, copper heat spreaders, and solder interconnects to minimize shear stress in power modules during thermal cycling.
- Design and qualification of DBC and AMB (active metal brazed) substrates for IGBT and SiC/GaN power semiconductor packages.
- Incoming material inspection and lot-to-lot statistical process control at ceramic substrate manufacturers.
- HTCC package development, where green-tape shrinkage data from TMA directly determines co-fired metallization pattern scaling factors.
- Failure analysis of delaminated metallization, cracked die-attach solder, and warped assemblies by comparing measured CTE with design assumptions.
- Finite-element thermal-stress simulation input, where temperature-dependent CTE curves from TMA improve joint-life predictions (e.g., Coffin–Manson solder fatigue models).
- Qualification of substrates for automotive (AEC-Q100/Q101 related reliability testing), aerospace, and downhole electronics where wide temperature excursions are expected.
FAQ
What is the typical CTE of alumina substrates?
96% alumina substrates typically exhibit a mean CTE of about 6.5–7.2 ppm/°C between 25 °C and 300 °C, rising to roughly 7.5–8.0 ppm/°C between 25 °C and 800 °C. High-purity 99.5%–99.6% grades generally measure slightly higher, around 7.0–8.2 ppm/°C over comparable intervals. Exact values depend on composition, microstructure, and measurement direction, which is why lot-specific TMA testing is recommended.
Which standards govern alumina substrate TMA testing?
The primary standards are ASTM E831 (TMA method for solid materials), ASTM E228 (push-rod dilatometry), ISO 11359-2 (TMA determination of coefficient of linear thermal expansion and glass transition temperature), and ISO 17561 (linear thermal expansion of fine ceramics by push-rod dilatometer). Related guidance for electronic assemblies may reference IPC and IEC test method series where substrate expansion is part of a broader qualification.
What temperature range and heating rate are used?
For as-fired substrates, the standard range is room temperature to 800–1000 °C at 3–5 °C/min under nitrogen or dry air. Green tape studies may extend to 1550–1600 °C in dilatometers to capture the full sintering curve. Faster rates are avoided because thermal gradients within the specimen bias the measured CTE.
How accurate is TMA for CTE measurement?
A properly calibrated TMA following ASTM E831 achieves CTE reproducibility on the order of ±2–5% for a material with a CTE of about 7 ppm/°C, with length-change resolution better than 100 nm. Push-rod dilatometry per ASTM E228 can improve absolute accuracy further and is preferred when certifying reference values.
Why does my specimen show a sudden slope change at high temperature?
In 96% alumina, the grain-boundary silicate glass phase begins to soften near 900–1000 °C, producing a sharp change in the expansion curve and possible probe penetration. This inflection is a real material characteristic, not instrument error, and defines the practical upper service temperature for the substrate.
Can metallized substrates be tested as received?
Yes. Screen-printed, plated, or DBC-copper bonded substrates can be tested directly to obtain composite expansion behavior; however, the reported CTE will reflect the layered structure rather than the pure ceramic. For ceramic-only data, the metallization should be removed by chemical etching or grinding before specimen preparation.