Substrate Glass Coefficient of Thermal Expansion Testing: Standards, Methods, and Applications

Substrate glass coefficient of thermal expansion (CTE) testing quantifies the dimensional stability of display and semiconductor glass substrates under thermal loading, which is critical for controlling thermal stress, warpage, and layer-to-layer compatibility in thin-film and packaging processes. This article explains the measurement principle, applicable standards such as ISO 7991, ASTM E228, ASTM E831, and GB/T 16920, typical sample preparation and testing procedures, and the main industrial applications of CTE characterization for substrate glass, as performed by accredited third-party laboratories.

Technical Overview

Substrate glass is the mechanical and dimensional foundation of thin-film transistor (TFT) arrays, color filters, touch sensors, and, increasingly, glass-core semiconductor packaging panels. Because device fabrication involves repeated thermal excursions — film deposition, activation and annealing, lithography bakes, and encapsulation — the linear coefficient of thermal expansion (CTE) of the substrate directly governs thermally induced stress, panel warpage, photolithographic overlay accuracy, and long-term device reliability.

The mean linear CTE over a temperature interval is defined as α = (L₂ − L₁) / [L₀ × (T₂ − T₁)], where L₀ is the specimen length at the reference temperature and L₁, L₂ are the lengths at temperatures T₁ and T₂. The result is expressed in ×10⁻⁶/K, commonly abbreviated as ppm/K.

Representative values at 20–300 °C are approximately 3.2–3.9 ×10⁻⁶/K for alkali-free aluminosilicate display glass, 3.25 ×10⁻⁶/K for borosilicate glass, and 8.5–9.5 ×10⁻⁶/K for soda-lime glass. Single-crystal silicon, the reference material in microelectronics, expands at about 2.6 ×10⁻⁶/K, so display substrates are engineered to remain close to silicon in order to limit film stress during high-temperature processing.

CTE testing of substrate glass is governed by internationally recognized methods published by standards organizations, principally ISO 7991, ASTM E228, ASTM E831, ASTM E289, and GB/T 16920. Accredited third-party laboratories routinely apply these methods for supplier qualification, incoming inspection, and research and development.

Testing Samples

Alkali-Free Display Substrate Glass

Coupons are cut from Generation 4.5 through 8.5 mother glass used for TFT-LCD and OLED panel production. Alkali-free aluminosilicate composition is the dominant display substrate type, specified for its low alkali content, high strain point, and low, tightly controlled CTE.

Specialty Glass for Microelectronics and Packaging

Specimens include borosilicate and alkali-free glass for glass-core packaging substrates with through-glass vias (TGV), wafer-level optics, MEMS caps and hermetic lids, sealing glass, and glass carriers used in temporary bonding processes.

Specimen Geometry and Preparation

• Push-rod dilatometry: rod or bar specimens typically 25–50 mm long (50 mm preferred) with a cross-section of approximately 5 × 5 mm or 5–7 mm diameter.
• End faces must be flat, mutually parallel, and perpendicular to the length axis, normally lapped or ground to within 0.01–0.02 mm.
• Thermomechanical analysis (TMA): specimens as short as 2–10 mm, useful when only limited material is available.
• Cut coupons must be annealed to relax residual machining stress, which otherwise relaxes during heating and distorts the low-temperature expansion data.
• Specimens are cleaned with solvent and lint-free wipes, and the sheet orientation (machine direction versus cross direction) is recorded when anisotropy is a concern.

Testing Items

Typical deliverables of a substrate glass CTE test program include:

• Mean linear CTE over specified intervals, most commonly 20–300 °C, with 20–100 °C, 20–200 °C, 30–380 °C, and 50–350 °C also frequently specified.
• Instantaneous (differential) CTE as a continuous function of temperature.
• Complete thermal expansion curve, ΔL/L₀ versus temperature, from room temperature up to 600–800 °C.
• Dilatometric glass transition temperature (Tg), identified at the inflection of the expansion curve.
• Dilatometric softening point, taken at the maximum of the expansion curve before viscous contraction begins.
• CTE anisotropy between machine direction and cross direction for float or rolled glass products.
• Expansion behavior under repeated thermal cycling, for stability and hysteresis assessment.

Testing Methods

Push-Rod Dilatometry (ISO 7991, ASTM E228, GB/T 16920)

The specimen is placed between a fused silica stage and a fused silica push rod inside a furnace. A small dead-load force maintains contact, and a linear variable differential transformer (LVDT) or capacitive sensor measures length change while a thermocouple positioned adjacent to the specimen records temperature. The system baseline is calibrated using certified reference materials traceable to national metrology institutes, such as fused silica, platinum, or borosilicate standards. Testing proceeds at a heating rate of 3–5 K/min (ISO 7991 specifies no more than 5 K/min), optionally in an inert atmosphere for high-temperature runs. The mean CTE is obtained by linear regression of the expansion curve over the specified temperature interval, with a well-controlled measurement uncertainty of approximately ±0.05 to ±0.1 ×10⁻⁶/K.

Thermomechanical Analysis (ASTM E831)

TMA uses a small specimen resting on a quartz platform under a low-force probe, typically 0.05–0.5 N. It is well suited to limited material quantities, sub-ambient temperature ranges down to about −120 °C, mapping of spatial CTE variation, and simultaneous determination of transition temperatures per ASTM E1545.

Interferometric and Optical Methods (ASTM E289)

Laser interferometric dilatometry measures expansion non-contact through optical fringe counting, achieving resolution on the order of 10⁻⁸/K. It is reserved for ultra-low-expansion glasses, certified reference work, and research applications rather than routine substrate qualification.

Data Reduction and Reporting

A complete test report documents specimen geometry, annealing condition, sheet orientation, heating rate, calibration reference, the exact temperature interval for each reported CTE value, the full expansion curve, dilatometric Tg and softening point, and the stated measurement uncertainty, ensuring comparability between laboratories and suppliers.

Applications

• TFT-LCD and OLED panel manufacturing: prediction of thermal stress and overlay error during a-Si, LTPS, and oxide semiconductor deposition, excimer-laser annealing, and furnace activation steps.
• Touch sensor and cover lamination: CTE matching of substrate, adhesive, and cover glass to minimize curl and delamination under thermal cycling.
• Advanced semiconductor packaging: glass-core substrates with through-glass vias, where tunable CTE between silicon (2.6 ×10⁻⁶/K) and organic laminates (14–17 ×10⁻⁶/K) controls warpage during reflow at peak temperatures near 260 °C.
• Wafer-level optics, MEMS, and hermetic packaging: dimensional compatibility of glass lids, caps, and sealing glass with metal and ceramic components.
• Quality control: incoming inspection, supplier qualification, and lot acceptance testing of mother glass at accredited third-party laboratories.
• Research and development: composition optimization of alkali-free glass to balance low CTE with high strain point and viscosity characteristics.

FAQ

Which temperature interval is used to report the CTE of display substrate glass?

The industry default is 20–300 °C (sometimes 25–300 °C). Some specifications use 30–380 °C or 50–350 °C to reflect specific process windows. Because CTE increases with temperature, values must only be compared over identical intervals.

How large must the sample be, and can finished panels be tested directly?

Rod specimens of 25–50 mm are cut from the panel; approximately 20 mm is the practical minimum length, as shorter specimens reduce displacement signal and degrade resolution. Full-size mother glass cannot be tested directly and must be sampled.

What measurement accuracy can be expected?

Repeatability within ±1–2% and an absolute uncertainty of roughly ±0.05–0.1 ×10⁻⁶/K are achievable with a calibrated push-rod dilatometer operated per ISO 7991 or ASTM E228.

Can the glass transition temperature be measured in the same test?

Yes. The dilatometric Tg is determined at the inflection point of the expansion curve, typically around 600–700 °C for alkali-free display glass. It may differ slightly from DSC or beam-bending values because each technique responds to different manifestations of the glass transition.

Why do specimens need to be annealed before testing?

Cutting and grinding leave residual stresses in the glass. When the specimen is heated, these stresses relax and produce apparent expansion anomalies at low temperature. Annealing near the annealing point followed by slow cooling establishes a stress-free baseline for accurate measurement.

Is the CTE of substrate glass direction-dependent?

Fusion-drawn substrates are essentially isotropic. Float or rolled glass can show a small difference between machine direction and cross direction, so directional specimens are tested when anisotropy is a specified concern.

What is the typical turnaround time?

A single dilatometer run, including calibration, takes 2–4 hours. Including specimen cutting, annealing, and reporting, accredited third-party laboratories typically deliver results within 5–10 working days.