Measurement Precision of the Coefficient of Thermal Expansion for PET Ceramic Substrates

The coefficient of thermal expansion (CTE) is a decisive parameter for PET ceramic substrates used in flexible electronics, thick-film circuits, and power device packaging, where even small CTE mismatches drive warpage, cracking, and delamination. Achieving high measurement precision depends on validated dilatometry or TMA procedures, certified reference materials, and a disciplined uncertainty analysis. This article reviews the sample types, tested parameters, applicable standards, achievable accuracy levels, and practical safeguards that define reliable CTE determination for PET ceramic substrates.

Technical Overview

PET ceramic substrates pair polyethylene terephthalate (PET) films or PET-based composites with functional ceramic layers such as alumina, aluminum nitride, silica, or ceramic-filled coatings. These constructions combine the flexibility, low weight, and low cost of polymer films with the thermal stability, barrier performance, and dielectric strength of ceramics, making them attractive for flexible printed circuits, membrane devices, heated panels, and hybrid electronic assemblies.

Thermal expansion behavior is the central design constraint for such laminates. PET expands at roughly 50 to 70 ppm/K below its glass transition, whereas dense engineering ceramics expand at only 3 to 8 ppm/K. During temperature excursions, this mismatch generates interfacial stresses that can crack sputtered or printed conductors, distort fine-pitch patterns, and separate layers. The mean linear CTE of the substrate, quoted in ppm/K over a defined temperature interval, must therefore be known with sufficient precision to feed warpage simulations, overlay budgets, and reliability models. Typical laboratory targets are a repeatability of 1 to 3 percent relative, or an absolute uncertainty on the order of 0.05 to 0.1 ppm/K for low-expansion, ceramic-rich specimens.

Testing Samples

Rigid Ceramic Substrates

High-purity alumina (96 percent and 99.6 percent Al2O3), aluminum nitride, silicon nitride, and zirconia-reinforced ceramic plates in thicknesses from 0.25 mm to 1.0 mm are tested as rectangular bars or disks sectioned to the dimensions required by the chosen method. Faces are ground flat and parallel within a few micrometers to eliminate probe seating errors.

PET Films and Ceramic-Coated PET

Biaxially oriented PET films from 50 micrometers to 250 micrometers, either bare or coated with sputtered alumina, sol-gel silica, or atomic-layer-deposited barrier layers, are conditioned at 23 degrees Celsius and 50 percent relative humidity before testing. Because oriented films are anisotropic, specimens are cut along both the machine direction and the transverse direction, and free-standing films may be stacked when a longer effective gauge length is required.

Laminated and Composite Specimens

Full laminate stacks, including adhesive layers, copper or printed conductors, and ceramic-filled PET cores, are tested in their as-manufactured state whenever the design value of interest is the CTE of the assembled construction rather than that of an individual layer. Specimens are taken from at least three positions across the panel to capture lot homogeneity.

Testing Items

A complete CTE characterization campaign for PET ceramic substrates typically reports the following quantities:

• Mean linear CTE over specified intervals, for example 23 to 100 degrees Celsius, 23 to 150 degrees Celsius, and 23 to 260 degrees Celsius, expressed in ppm/K.
• Instantaneous (differential) CTE as a continuous function of temperature, computed from the first derivative of the expansion curve.
• Glass transition temperature of the PET phase, identified from the characteristic step increase in CTE, together with the magnitude of the step.
• Anisotropy of expansion, including machine direction versus transverse direction for films and in-plane versus through-thickness behavior for laminates.
• Hysteresis between heating and cooling curves, reported as the maximum deviation in strain at matched temperatures.
• Residual dimensional change after thermal cycling, indicating permanent shrinkage or stress relaxation of the PET layer.

Testing Methods

Thermomechanical Analysis (TMA)

TMA per ASTM E831 and ISO 11359-2 is the workhorse for thin substrates. A flat silica probe applies a small static force, typically 0.02 to 0.05 N, to a 2 to 10 mm specimen, and length change is recorded with a resolution near 0.1 micrometers at heating rates of 5 degrees Celsius per minute. With calibration against certified reference materials, TMA delivers CTE repeatability of 1 to 3 percent for rigid ceramics and 0.2 to 0.5 ppm/K for polymer-dominated laminates, which is sufficient for most qualification and screening purposes.

Push-Rod Dilatometry

Horizontal or vertical push-rod dilatometers per ASTM E228 accept longer specimens, usually 20 to 50 mm, and provide displacement resolution in the 10 to 50 nanometer range. The longer gauge length improves sensitivity for low-expansion ceramics, allowing combined expanded uncertainties of roughly 1 to 2 percent relative, or 0.05 ppm/K absolute, on well-prepared alumina and aluminum nitride bars. A fused-silica or alumina reference run in the same cycle is used to correct the system baseline.

Laser Interferometry and Optical Extensometry

Interferometric dilatometry per ASTM E289 measures absolute expansion with sub-nanometer resolution and is reserved for ultra-low-CTE ceramics or arbitration measurements. Non-contact optical methods, including video extensometry and digital image correlation, suit anisotropic PET films and patterned substrates where probe contact would dent or stress the polymer layer, achieving typical uncertainties of 2 to 5 percent while recording both in-plane axes simultaneously.

Key Precision Factors

• System calibration with certified reference materials such as fused silica (NIST SRM 739), sapphire, and platinum over the full working temperature range.
• Temperature measurement and uniformity, with calibrated thermocouples placed close to the specimen and furnace gradients held within 1 degree Celsius.
• Specimen geometry, particularly length accuracy of 5 micrometers and end-face parallelism, which scale directly into the CTE error budget.
• Probe force kept at the minimum stable value to avoid indenting PET-rich layers and to suppress creep artifacts.
• Heating rates of 5 degrees Celsius per minute or lower to minimize thermal lag between specimen and sensor.
• Inert, dry atmosphere such as nitrogen or argon to prevent oxidation of conductors and moisture-driven swelling of the PET phase.
• Specimen conditioning to ISO 291 (23 degrees Celsius, 50 percent relative humidity for at least 24 hours) and replicate runs, usually three heating cycles, with the first cycle discarded when conditioning history matters.

A disciplined uncertainty budget that combines displacement resolution, length measurement, temperature calibration, gradient effects, and repeatability typically yields expanded uncertainties (coverage factor k = 2) of 1 to 2 percent for dilatometry on rigid ceramics and 3 to 5 percent for thin, polymer-dominated laminates.

Applications

• Flexible and printed electronics, where substrate CTE controls overlay registration and conductor integrity across reflow and service temperatures.
• Thick-film heaters, sensors, and membrane switches built on ceramic-coated PET platforms.
• Power electronics packaging, where ceramic substrate CTE is matched to silicon and silicon carbide dies and to copper metallization to limit solder joint fatigue.
• Finite element warpage and thermo-mechanical stress simulations that require temperature-dependent CTE curves as validated inputs.
• Incoming inspection and supplier qualification, including lot-to-lot statistical comparison against specification limits.
• Research and development of ceramic-filled PET composites engineered toward tailored, application-specific expansion behavior.

FAQ

Question: What measurement precision can realistically be achieved for PET ceramic substrate CTE?

Answer: With calibrated push-rod dilatometry, expanded uncertainties of about 1 to 2 percent relative, or roughly 0.05 ppm/K on low-expansion ceramics, are achievable. TMA on thin laminates typically delivers repeatability of 1 to 3 percent, while optical methods on films reach 2 to 5 percent depending on gauge length and surface pattern quality.

Question: Which standard should govern the test?

Answer: ASTM E831 or ISO 11359-2 covers TMA on plastics and composites, ASTM E228 governs push-rod dilatometry of solid materials, and ASTM E289 addresses interferometric measurements for the highest accuracy requirements.

Question: What specimen dimensions are required?

Answer: TMA accepts 2 to 10 mm lengths with thickness from 0.1 mm upward; dilatometry prefers 20 to 50 mm bars with ground, parallel faces; interferometry typically uses 5 to 20 mm specimens with polished optical faces. Film specimens are stacked when a longer gauge length is needed.

Question: What temperature range is typical?

Answer: Most PET ceramic substrate evaluations run from room temperature to 150 or 200 degrees Celsius, staying below the melting range of PET while capturing the glass transition step. Ceramic-only substrates are often characterized up to 300 or 500 degrees Celsius when packaging processes require it.

Question: Why can TMA and dilatometry results differ for the same material?

Answer: Differences arise from gauge length, probe force effects on soft layers, heating rate lag, and the calibration chain. Agreement within the combined uncertainty of both methods should be demonstrated, and arbitration measurements are best performed by interferometry or verified through interlaboratory comparison.