PET Ceramic Substrate Coefficient of Thermal Expansion (CTE) Testing Accuracy: Methods, Standards, a

The coefficient of thermal expansion (CTE) is a decisive reliability parameter for PET ceramic substrates, governing dimensional stability, interfacial stress, warpage behavior, and solder joint fatigue life in electronic packaging and thermal management applications. Achieving high testing accuracy requires rigorous control of instrumentation, sample preparation, environmental conditioning, and data regression methodology. This article presents a comprehensive technical review of CTE testing accuracy for PET ceramic substrates, covering measurement principles, international standards, uncertainty sources, and best practices for reproducible characterization.

Technical Overview

PET ceramic substrates are composite dielectric and structural materials in which ceramic fillers or ceramic layers are combined with a polyethylene terephthalate (PET) matrix or carrier film. They are widely used in flexible printed circuits, film-based thermal management modules, LED substrate assemblies, and microelectronic packaging where low cost, flexibility, and controlled dimensional behavior are simultaneously required. Because these substrates are laminated or soldered to materials with markedly different expansion characteristics, the coefficient of thermal expansion (CTE) must be known with high accuracy to predict interfacial stress and long-term reliability.

The linear CTE, typically expressed in ppm/K, quantifies the fractional length change of the material per unit temperature change. For PET-based systems, the CTE is inherently anisotropic and temperature-dependent: below the glass transition (Tg, approximately 70–80 °C for PET), filled substrate CTE values commonly fall between 10 and 30 ppm/K, while above Tg the polymer matrix softens and the effective expansion rate can rise sharply. In contrast, adjacent engineering materials such as copper foil expand at about 17 ppm/K, alumina ceramic at 6.5–7.2 ppm/K, aluminum nitride at approximately 4.5 ppm/K, and silicon at 2.6 ppm/K. Accurately resolving these differences, and the transitions within the substrate itself, demands measurement precision on the order of ±0.1 to ±0.5 ppm/K.

• Definition of linear CTE and its temperature-dependent behavior in polymer-ceramic composites
• Typical CTE ranges of PET ceramic substrates versus copper, silicon, and monolithic ceramics
• Accuracy targets required for packaging design, warpage simulation, and qualification testing
• Influence of anisotropy, Tg transitions, moisture, and residual stress on measurement validity

Testing Samples

Ceramic-Filled PET Dielectric Films

Thin composite films, generally 25 to 250 micrometers in thickness, are tested as free-standing strips or as laminated multi-layer stacks. Because individual films may be too thin for reliable push-rod contact, several plies are often stacked and thermally bonded to produce a specimen with sufficient stiffness, typically achieving a combined thickness of 0.5 mm or greater while preserving through-thickness representativeness.

Rigid Ceramic Substrate Laminates on PET Carriers

Sandwich constructions consisting of a ceramic layer, such as alumina or a ceramic-filled coating, bonded to a PET film or core are tested in the as-laminated state to capture the true composite expansion behavior. Specimens must retain all functional layers, including adhesives and barrier coatings, since thin interlayers can measurably shift the apparent CTE of the assembly.

Sample Geometry and Quantity

For thermomechanical analysis, specimens measuring 3 to 10 mm in width and 5 to 10 mm in length are standard, with thickness between 0.5 and 3 mm preferred for stable probe contact. Push-rod dilatometry accommodates longer specimens of 25 to 50 mm, which reduces relative length-measurement error. Ends must be ground flat, parallel, and perpendicular to the measurement axis within approximately 10 micrometers. A minimum of three specimens per direction is recommended so that mean values and standard deviations can be reported.

Conditioning and Preparation

PET is hygroscopic and moisture-sensitive; absorbed water and its subsequent loss during heating can be misinterpreted as expansion or shrinkage. Specimens are conditioned at 23 ± 2 °C and 50 ± 5 % relative humidity for a minimum of 24 hours prior to testing. When a dry-state CTE is required, specimens are pre-baked at 105 °C or below the Tg until mass stabilization. Edges are deburred, and any pre-test annealing is documented, since residual orientation stress relaxation in PET can introduce irreversible shrinkage unrelated to true thermal expansion.

Testing Items

A complete CTE characterization program for PET ceramic substrates covers several interrelated measurands, each with distinct accuracy considerations:

• Linear CTE below the glass transition (alpha 1), typically regressed between 30 and 60 °C
• Linear CTE above the glass transition (alpha 2), regressed over a window such as 100 to 150 °C
• Glass transition temperature determined from the intersection of the two expansion tangents
• Total percent expansion across the operational range, for example from 25 to 150 °C
• Z-axis or through-thickness CTE, which governs plated-hole reliability and via cracking
• Reversible versus irreversible dimensional change, separating elastic expansion from stress relaxation, moisture loss, and crystallization effects

Each item requires an explicit definition of the regression interval, heating rate, and specimen orientation, because these parameters directly influence the reported accuracy. Omitting them is a frequent source of inter-laboratory discrepancy.

Testing Methods

Thermomechanical Analysis (TMA)

TMA is the most widely applied technique for PET ceramic substrates and is governed by ASTM E831, ISO 11359-2, and IPC-TM-650 method 2.4.24 for laminate qualification. A specimen rests on a fused silica platform and is loaded by a low-force expansion probe, typically below 0.1 N, while length change is recorded by an LVDT or capacitive sensor with displacement resolution of 0.01 to 0.1 micrometer. Heating rates of 3 to 5 °C/min in nitrogen minimize thermal lag and oxidative degradation. With proper calibration, TMA delivers CTE repeatability of approximately ±2 to 5 % or ±0.3 to 0.5 ppm/K, whichever is greater, which is adequate for most design and qualification purposes.

Push-Rod Dilatometry

Vertical or horizontal push-rod dilatometry per ASTM E228 uses longer specimens and a fused silica reference enclosure, improving length measurement statistics. The technique suits thicker, more rigid substrate laminates and achieves repeatability near ±1 to 3 %. Corrections for the expansion of the apparatus itself are made by running a baseline with a certified reference such as fused silica or sapphire. Slightly higher contact forces are used than in TMA, so probe indentation must be verified as negligible, particularly above the PET Tg where the matrix creeps.

Laser Interferometric Dilatometry

For applications demanding the highest accuracy, such as substrate qualification for precision optical or metrology-grade assemblies, absolute interferometric dilatometry per ASTM E289 measures displacement optically with nanometer resolution. This method achieves uncertainties better than ±0.05 ppm/K and is capable of characterizing low-expansion ceramic layers and near-zero differential expansion behavior, though it requires polished, optically parallel specimen faces and more elaborate fixturing.

Calibration and Accuracy Verification

Temperature sensors are calibrated at the melting points of indium, tin, and zinc or against certified thermocouples, targeting uncertainty within ±0.5 to 1 °C. Displacement channels are verified using certified gauge blocks or standard reference materials spanning known CTE values: fused silica near 0.5 ppm/K, sapphire near 5.3 ppm/K, and aluminum near 23 ppm/K. A fused silica baseline run immediately before specimen testing captures and subtracts system expansion, and periodic verification ensures the instrument remains within its stated uncertainty budget.

Factors Affecting Measurement Accuracy

• Temperature calibration error and thermal lag between the sensor and the specimen, aggravated by excessive heating rates
• Displacement resolution, probe force, and indentation or creep of the softened PET phase above Tg
• Specimen length measurement uncertainty, which enters directly into the CTE denominator
• Moisture desorption shrinkage that biases the apparent expansion slope
• Relaxation of residual orientation stress in PET, producing irreversible dimensional change
• Anisotropy between machine direction and transverse direction, and between in-plane and Z-axis behavior
• Poor flatness or parallelism of specimen ends, causing contact instability and false zero shifts

Best practice combines slow or stepwise heating, low probe force, pre-test conditioning with documented humidity history, baseline subtraction, well-defined tangent regression windows that avoid transition zones, and replicate testing with full reporting of mean values, standard deviations, and all test parameters.

Applications

Accurate CTE data for PET ceramic substrates underpins nearly every stage of electronic product development and manufacturing control:

• Flexible printed circuit and coverlay design, where matching the CTE of copper foil and adhesive systems limits curl and registration error
• Chip-on-film and direct chip attach, where minimizing the CTE gap to silicon at 2.6 ppm/K reduces bump and adhesive joint stress
• LED and power module substrates, in which the ceramic phase controls heat spreading while the PET layer must not introduce differential warpage
• Finite element warpage and thermomechanical simulation, where input uncertainty propagates directly into predicted stress and fatigue life
• Solder joint reliability assessment using strain-range models such as Coffin-Manson, which are highly sensitive to differential expansion inputs
• Incoming inspection, supplier qualification, and lot-to-lot statistical process control for substrate manufacturers
• Lamination process optimization, where CTE mismatch correlates with delamination risk during thermal cycling

FAQ

What measurement accuracy is realistically achievable for PET ceramic substrates?

With a properly calibrated TMA following ASTM E831, CTE repeatability of ±0.3 to 0.5 ppm/K or roughly ±3 to 5 % relative is typical for filled PET composites. Push-rod dilatometry per ASTM E228 improves this to approximately ±1 to 3 %, while laser interferometry per ASTM E289 can reach uncertainties below ±0.05 ppm/K when specimen geometry and surface quality permit.

Which standard should be selected for testing?

ASTM E831 or ISO 11359-2 is appropriate for routine TMA measurement of polymer-based substrates; IPC-TM-650 method 2.4.24 is the recognized procedure for copper-clad laminate qualification including Z-axis CTE and Tg; ASTM E228 suits thicker rigid laminates; and ASTM E289 is reserved for highest-accuracy or low-expansion requirements. The chosen standard and all deviations should be stated in the test report.

Why do machine-direction and transverse-direction CTE results differ?

PET films are biaxially oriented during manufacturing, and ceramic fillers can align during casting or calendering. The resulting molecular and particulate orientation produces in-plane anisotropy that frequently differs by several ppm/K, so both directions must be tested and reported separately.

How does moisture affect the measured CTE?

Absorbed moisture in PET can desorb during the heating ramp, superimposing shrinkage on true thermal expansion and biasing the CTE slope by several ppm/K. Standard conditioning at 23 ± 2 °C and 50 ± 5 % relative humidity, or a documented dry-state pre-bake, is essential, and a second heating run can be used to distinguish reversible expansion from drying effects.

What temperature range is valid for testing PET-based substrates?

Testing normally spans 25 to 150 °C or up to about 200 °C, staying safely below the PET melting range of 250 to 260 °C. Regression windows should exclude the Tg transition region and any relaxation events, with alpha 1 typically taken from 30 to 60 °C and alpha 2 from 100 to 150 °C, unless the material specification defines otherwise.

How many specimens are required for a defensible accuracy statement?

A minimum of three specimens per orientation is standard practice. Reporting the mean, standard deviation, regression interval, heating rate, conditioning history, and calibration verification together constitutes a complete and auditable accuracy statement for PET ceramic substrate CTE data.