Substrate glass crystallization performance testing evaluates the devitrification resistance of glass substrates used in flat-panel displays, semiconductor packaging, photovoltaics, and other high-precision electronic applications. By measuring key thermal parameters such as liquidus temperature, crystallization onset temperature, and crystallized phase composition, manufacturers can define safe process windows, prevent crystal-related defects, and ensure product reliability. Accredited third-party laboratories perform these tests in accordance with internationally recognized ISO, ASTM, and GB standards to support quality control, failure analysis, and new composition development.
Technical Overview
Substrate glass is a high-precision flat glass that serves as the structural and functional carrier for thin-film transistors, color filters, touch sensors, semiconductor interposers, and photovoltaic cells. It is typically produced by the fusion draw, float, or slot draw process, during which molten glass is shaped into sheets at high temperature and then cooled rapidly through the glass transition range to preserve an amorphous structure.
From a thermodynamic perspective, glass is a metastable supercooled liquid. When glass is held for a sufficient time within the temperature interval between the glass transition temperature and the liquidus temperature, nucleation and crystal growth—collectively known as crystallization or devitrification—can occur. In continuous forming operations such as fusion draw, the molten ribbon resides for an extended period at viscosities between approximately 103 and 105 Pa·s. If the local glass temperature approaches or falls below the liquidus temperature, crystals nucleate on refractory surfaces, in the bulk, or at the glass surface, leading to haze, nodules, inclusions, streak defects, sheet breakage, and reduced production yield.
Crystallization performance testing therefore quantifies how resistant a given glass composition is to devitrification under realistic thermal histories. The core metrics include the liquidus temperature, the crystallization onset and peak temperatures measured by differential scanning calorimetry, the glass transition temperature, the thermal stability interval, the identity and amount of the crystalline phases that form, and the crystallization kinetics parameters. Together, these data allow engineers to verify that an adequate temperature margin exists between the forming window and the crystallization region, and to qualify substrate glass for demanding downstream thermal processes such as ion exchange, annealing, laser cutting, and through-glass via formation.
Testing Samples
Alkali-Free Aluminosilicate Display Glass Coupons
Coupons cut from production sheets of alkali-free aluminosilicate or aluminoborosilicate glass, the dominant substrate families for TFT-LCD and OLED applications. Typical coupon sizes range from 10 × 10 mm to 25 × 25 mm, with sheet thicknesses from 0.1 mm to 1.1 mm. Edges are ground and surfaces are cleaned to eliminate handling debris that could seed false nucleation sites.
Borosilicate, Soda-Lime, and Specialty Glass Substrates
Samples of borosilicate glass for semiconductor and packaging applications, soda-lime float glass for photovoltaic modules, and specialty glasses such as high-silica or rare-earth-doped compositions, evaluated with the same crystallization test protocols after appropriate parameter adaptation.
Glass Powder and Frit Specimens
For differential scanning calorimetry and differential thermal analysis, bulk glass is crushed and sieved into a defined particle size fraction, commonly below 45 μm or a specified 45–100 μm band. Powdered samples of approximately 10–50 mg maximize the surface-nucleated crystallization signal and improve thermal contact in the crucible. Powder specimens are also used to evaluate sealing frits and glass powders intended for sintering or vitrified-bond applications.
Melt Cullet and Process Intermediates
Cullet, melter overflow samples, and beads drawn directly from the production line can be tested to monitor whether upstream melting conditions are shifting the effective crystallization behavior, for example through volatilization of boron or alkali components.
Thermally Treated Reference Specimens
Coupons that have been heat-treated at selected temperatures and hold times are prepared as reference specimens for phase identification, crystal morphology characterization, and quantitative image analysis of the devitrified layer.
Testing Items
Depending on the engineering objective, one or more of the following parameters are reported:
- Liquidus temperature (TL): the lowest temperature at which the glass melt remains in equilibrium with its primary crystalline phase, beyond which complete melting occurs.
- Crystallization onset temperature (Tx): the temperature at which exothermic crystal formation begins during a controlled heating scan.
- Crystallization peak temperature (Tp): the temperature of maximum crystallization rate, corresponding to the apex of the DSC exotherm.
- Glass transition temperature (Tg): the temperature marking the transformation between the glassy and supercooled liquid states, determined as the midpoint or fictive-related inflection of the heat-flow step.
- Thermal stability parameter (ΔT = Tx − Tg): an indicator of the resistance of the glass to devitrification upon reheating; larger values denote greater stability.
- Forming process margin: the difference between the temperature corresponding to the working viscosity and the liquidus temperature, a critical design criterion for fusion draw and float processes.
- Crystallized phase identification: determination of the crystal species formed, such as wollastonite, mullite, cristobalite, spinel, or nepheline-type phases, by X-ray diffraction.
- Crystallized fraction and crystal density: the weight or volume percentage of crystals and the number of crystals per unit area after a defined heat treatment.
- Crystal morphology and size distribution: characterization of dendritic, acicular, granular, or layered growth patterns and their statistical size distribution.
- Crystallization kinetics parameters: the effective activation energy for crystallization, the Avrami exponent, and the rate constant, derived from multi-rate thermal analysis or isothermal treatment series.
- Time–temperature–transformation (TTT) behavior: the nose curve describing the time required for a detectable crystallized fraction at each holding temperature.
- Surface versus bulk nucleation tendency: assessment of whether devitrification initiates preferentially at surfaces or within the volume, which governs defect risk in finished substrates.
Testing Methods
Liquidus Temperature by the Gradient Furnace Method
Following the practices described in ASTM C829, specimens are positioned along a furnace with a controlled, calibrated linear temperature gradient and held for a defined soaking period. After rapid cooling, each specimen is examined with an optical microscope in transmitted and reflected light to detect the presence of crystals. Trials bracket the temperature at which crystals first appear, and the liquidus temperature is reported with a stated uncertainty. Platinum or ceramic trays are used to avoid container-induced contamination, and duplicate runs confirm repeatability.
Differential Scanning Calorimetry and Differential Thermal Analysis
DSC or DTA measurements are performed on powdered or bulk specimens in alumina or platinum crucibles under a dry nitrogen purge, typically at heating rates of 5, 10, 15, and 20 K/min. The test methods of ASTM E1356 and ASTM E794 are applied to assign the glass transition, crystallization onset, crystallization peak, and any subsequent melting events. Exothermic peak areas are integrated to estimate the enthalpy and relative extent of crystallization.
Multi-Rate Kinetic Analysis
Crystallization peak temperatures obtained at several heating rates are analyzed with the Kissinger, Ozawa, and Augis–Bennett relations to extract the effective activation energy. Isothermal transformation data are fitted to the Johnson–Mehl–Avrami–Kolmogorov model to obtain the Avrami exponent, which distinguishes surface-controlled from bulk-controlled crystallization mechanisms and supports lifetime extrapolations for process design.
Isothermal Devitrification Heat Treatment
Coupons are held in precision tube or muffle furnaces at a matrix of temperatures spanning the expected crystallization range, with hold times from minutes to tens of hours, followed by rapid quenching. The heat-treated specimens are then inspected to count surface and internal crystals, measure the depth of the devitrified layer, and construct TTT curves that define safe residence times at each temperature.
X-ray Diffraction Phase Analysis
Heat-treated and as-received samples are scanned by X-ray diffraction to identify the crystalline phases and to quantify the crystallized fraction using reference-intensity-ratio or full-pattern Rietveld refinement against an amorphous internal standard. This confirms whether the crystals observed microscopically match the thermodynamically predicted primary phase.
Microscopy and Quantitative Image Analysis
Optical microscopy, polarized-light microscopy, and scanning electron microscopy with energy-dispersive spectroscopy are used to document crystal morphology, elemental composition, and spatial distribution. Digital image analysis converts micrographs into statistical data on crystal number density, equivalent diameter, and area fraction, providing quantitative acceptance metrics for production control.
Hot-Stage Microscopy
A controlled-atmosphere hot stage mounted on a microscope allows direct, continuous observation of crystal nucleation and growth in real time, which is particularly useful for distinguishing the initiation sites and growth habits of devitrification in surface-sensitive applications.
Applications
Crystallization performance testing of substrate glass supports decision-making across the entire product life cycle:
- Display substrate qualification for TFT-LCD and OLED panels, where even trace devitrification causes visible defects and yield loss in high-generation production lines.
- Forming process design for fusion draw, float, and slot draw lines, ensuring an adequate viscosity–liquidus margin so that the ribbon never enters the crystallization regime.
- New glass composition development, using ΔT, liquidus temperature, and kinetics data to rank candidate alkali-free aluminosilicate and aluminoborosilicate formulations.
- Downstream thermal process validation for ion-exchange strengthening, thermal annealing, laser processing, and packaging reflow, confirming that process temperatures and residence times remain outside the TTT nose region.
- Semiconductor packaging substrates and through-glass via carriers, where thermal stability during via metallization and bonding is essential.
- Photovoltaic front and back glass, glass fiber production, and sealing glass evaluation, where sustained high-temperature exposure makes devitrification resistance a primary durability criterion.
- Failure analysis and supplier comparison, using phase identification and crystal morphology to trace the origin of inclusions found in finished substrates.
FAQ
What is the liquidus temperature and why is it critical for substrate glass?
The liquidus temperature is the lowest temperature at which the glass melt coexists in equilibrium with its primary crystalline phase. Above it, the glass is fully molten and thermodynamically stable against crystallization. In forming processes, the glass must be shaped at temperatures comfortably above the liquidus while still viscous enough to be drawn into a sheet; the size of this margin directly determines process stability and defect risk.
What does a larger ΔT (Tx − Tg) indicate?
A larger ΔT indicates a wider temperature interval between the glass transition and the onset of crystallization, meaning the glass can be reheated, annealed, or thermally processed with a lower risk of devitrification. It is one of the most widely used comparative indicators of glass stability against crystallization.
What is the difference between surface and bulk crystallization?
Surface crystallization initiates at external surfaces or interfaces, often driven by surface energy, abrasion, or contamination, whereas bulk crystallization nucleates homogeneously or on internal heterogeneities within the volume. Powder-based DSC testing emphasizes surface crystallization, while heat-treated bulk coupons and microscopy reveal the dominant mechanism in the actual product geometry.
Which standards govern these tests?
Liquidus temperature determination follows the gradient-method practices of ASTM C829; glass transition and crystallization event assignment by thermal analysis follow ASTM E1356 and ASTM E794; and complementary national methods in the GB/T series and ISO thermal analysis conventions are applied where appropriate. Testing is performed in accredited third-party laboratories to ensure traceable calibration and defensible reporting.
What sample forms and quantities are required?
Bulk coupons of roughly 10 × 10 mm to 25 × 25 mm are sufficient for gradient furnace and heat-treatment studies, while DSC/DTA requires only 10–50 mg of powder from a defined sieve fraction. For kinetics studies, several specimens per heating rate or per isothermal condition are recommended to obtain statistically robust parameters.
How are results used to set production process windows?
The liquidus temperature, TTT curves, and kinetics parameters are combined with viscosity data to define the maximum allowable residence time at each forming temperature. Process engineers then set throughput, melter residence, and isopipe conditions so that the entire thermal path of the glass remains outside the crystallization region with a documented safety margin.