Determination of Bubble Content in Substrate Glass: Testing Methods, Standards, and Quality Control

Bubble content determination is a critical quality control test for substrate glass used in flat panel displays, touch panels, optical devices, and semiconductor packaging, where even microscopic gaseous inclusions can produce visible optical defects and reduce panel yield. This technical article explains how bubbles form in substrate glass, which sample types are tested, the key measurement parameters, standard testing methods such as optical microscopy and automated optical inspection, and how results are evaluated against industry specifications. It is a practical reference for glass manufacturers, display engineers, and quality professionals who require accurate and repeatable bubble measurement.

Technical Overview

Substrate glass is a high-quality flat glass sheet that serves as the mechanical and functional carrier for thin-film transistors, color filters, touch sensors, and OLED device layers. For display applications it is typically an alkali-free boroaluminosilicate glass with a thickness between 0.1 mm and 1.1 mm, produced by fusion draw or float processes. Because subsequent device layers are deposited directly on the glass surface, the internal quality of the sheet must approach optical-grade homogeneity.

Bubbles are gaseous inclusions trapped in the glass during melting and refining. They typically contain residual fining gases such as carbon dioxide, sulfur dioxide, oxygen, nitrogen, or water vapor, and they can range in size from a few micrometers to several hundred micrometers. Incomplete fining, reboil phenomena, refractory corrosion, unfavorable raw-material particle size distributions, and furnace atmosphere fluctuations are the most common formation mechanisms.

Even very small bubbles degrade product quality in several ways:

• Optical scattering and refraction that create visible dot defects or mura on finished panels
• Light leakage in black-matrix regions of LCD modules
• Local stress concentration that reduces mechanical strength and increases fracture risk during cutting, thinning, or module assembly
• Interference with photolithography, thin-film deposition, and laser processes on the production line
• Potential device failure when a bubble is located close to active device layers

Bubble content determination therefore quantifies the number, size, and spatial distribution of bubbles within a defined inspection area or volume and compares the results with specification limits, usually expressed as size classes. Relevant reference frameworks include GB 11614 for the classification of bubble defects in flat glass, ASTM C1036 for flat glass quality requirements, ISO 10110-3 for the indication of bubbles and inclusions on optical drawings, and JIS R 3201 for flat glass evaluation. Display-grade substrate specifications are typically far stricter than general flat glass grades, with detection thresholds commonly set in the range of 20 to 50 micrometers.

Testing Samples

Display Substrate Glass (Mother Glass)

Alkali-free glass sheets for TFT-LCD and OLED applications, sampled from production generations ranging from small substrates up to Generation 8.5 and larger mother glass. Depending on the purpose, either full sheets are inspected or coupons of approximately 100 mm by 100 mm are cut from designated positions such as the center, corners, and edges of the sheet in order to assess distribution uniformity.

Cover Glass and Touch Panel Glass

Aluminosilicate or soda-lime blanks intended for chemical strengthening and subsequent use as cover windows. Bubble inspection is performed before strengthening, since bubbles above a critical size can act as fracture origins during ion exchange or impact loading.

Float and Soda-Lime Glass Substrates

Glass used in photovoltaic modules, appliance panels, and electronic assemblies, where bubble content affects both appearance grades and downstream processing. Sampling follows the lot classification rules of the applicable flat glass standard.

Glass Wafers and Specialty Substrates

Thin glass wafers and micro-sheets for semiconductor packaging, microfluidics, and optical components. These are inspected under tighter size classes because downstream feature dimensions are comparable to the bubble sizes of interest.

Testing Items

The core measurable quantities in bubble content determination are listed below. All of them are reported together with the inspection area, magnification, and detection threshold of the method used.

• Bubble number density, expressed as count per cm², per dm², or per defined inspection zone
• Individual bubble size, reported as the maximum dimension or the equivalent circular diameter in micrometers
• Size class distribution, for example bands such as below 50 µm, 50 to 100 µm, 100 to 200 µm, and above 200 µm
• Maximum single bubble size observed in the inspection area
• Cumulative defect area or projected area fraction occupied by bubbles
• Spatial distribution map showing bubble coordinates and any clustering tendency
• Bubble depth through the sheet thickness, when relevant for device integration
• Simultaneous count of solid inclusions and stones, which are frequently evaluated in the same inspection

Testing Methods

Microscopic Inspection Method

A transmitted-light microscope operated in bright-field or dark-field mode examines the sample at magnifications of 50× to 500×. The sample is mounted on a translation stage and the inspection area is scanned in a defined grid pattern while focusing through the sheet thickness at successive depth planes so that bubbles at any depth are detected. Bubble dimensions are measured with a calibrated eyepiece reticle or, more commonly, with a digital camera and image analysis software that identifies bubbles by contrast and shape. The method detects bubbles down to approximately 5 to 10 µm and is the reference technique for small coupons and method validation.

Collimated Light (Parallel Light) Method

The sample is placed in an expanded collimated beam and the transmitted light is projected onto a screen or camera. Bubbles refract and scatter the parallel light, appearing as sharply defined dark spots against a bright background. The technique covers large areas rapidly, which makes it suitable for full-sheet screening and for flat glass grading in accordance with GB 11614 and JIS R 3201. Typical detection limits are in the range of 30 to 50 µm, depending on beam quality and camera resolution.

Laser Scattering and Oblique Illumination Method

A scanning laser beam or an oblique sheet of light passes through the glass, and scattered light is collected by a photodiode or camera. Because even sub-micron discontinuities scatter light strongly at gas-glass interfaces, this approach achieves the highest sensitivity, detecting bubbles smaller than 5 µm. Confocal or triangulation configurations can additionally provide depth information, which distinguishes internal bubbles from surface contamination.

Automated Optical Inspection (AOI)

Machine-vision systems combining telecentric optics, line-scan cameras, and LED transmission illumination inspect complete Generation-size sheets at production speed. Software algorithms classify each detected defect by equivalent diameter and position, separating bubbles from scratches, inclusions, and surface particles. AOI is the standard solution for routine lot release in substrate glass manufacturing, and its results are periodically cross-verified against microscopic measurements.

General Test Procedure

• Define the sampling plan: number of sheets or coupons per production lot and the sampling positions on each sheet
• Clean the sample surfaces with isopropyl alcohol and lint-free wipes, then dry in a dust-free environment
• Mark the inspection area with an appropriate edge exclusion zone, typically 10 mm, to avoid cutting-induced artifacts
• Calibrate the system magnification with a certified stage micrometer and confirm the detection threshold using reference artifacts
• Scan the inspection area continuously or on a grid, focusing through the full thickness to capture bubbles at all depths
• Record coordinates and dimensions, classify each bubble into the specified size band, and distinguish internal bubbles from surface defects by through-focus behavior
• Calculate number density and distribution statistics, compare with acceptance criteria, and issue the test report

Typical acceptance criteria for display substrate glass require, for example, that no bubble exceed a defined maximum diameter anywhere on the sheet, that only a limited number of bubbles fall within an intermediate size band, and that bubbles below the counting threshold, often 50 µm, are disregarded. The exact limits are set by the customer technical specification or the applicable flat glass standard.

Applications

Bubble content determination supports the entire life cycle of substrate glass, from melt development to panel production:

• Melting and fining process monitoring, where bubble trends reveal furnace atmosphere or batch-related problems early
• Lot release inspection and product certification for glass manufacturers
• Incoming material verification at display panel and module fabrication plants
• Research and development of new glass compositions and melting trials
• Root cause analysis of visible panel defects by correlating device-level dot defects with glass bubble maps
• Supplier qualification and periodic quality audits performed by accredited third-party laboratories

In all of these scenarios, consistent measurement conditions and traceable calibration are essential, because acceptance decisions hinge on counts within narrowly defined size classes.

FAQ

What is the difference between a bubble and an inclusion?

A bubble is a gaseous void that appears as a bright, high-contrast, roughly circular or elliptical feature under transmitted light, whereas an inclusion or stone is a solid particle such as a refractory fragment or unmelted batch material with an irregular, opaque appearance. Both defect types are usually evaluated in the same inspection, and ISO 10110-3 provides a unified size-class notation for both on optical drawings.

What is the smallest bubble that can be detected?

The detection limit depends on the method. Automated optical inspection typically resolves bubbles of about 20 to 30 µm, the collimated light method about 30 to 50 µm, optical microscopy about 5 to 10 µm, and laser scattering methods can detect bubbles below 5 µm. The report must always state the detection threshold of the method used.

Which standards govern bubble content evaluation?

GB 11614 defines size classes and allowable bubble counts for flat glass, ASTM C1036 specifies flat glass quality including bubble and void limits, ISO 10110-3 defines the indication of bubbles and inclusions for optical elements, and JIS R 3201 covers corresponding Japanese flat glass requirements. Display-grade substrates are additionally governed by customer technical specifications that are considerably stricter than general flat glass grades.

How are results reported?

A complete report identifies the sample and sampling position, the test method and magnification, the inspected area, the detection threshold, the total bubble count broken down by size class, the number density, the maximum observed bubble size, a distribution map where required, and a clear pass or fail conclusion against the applicable criteria.

Can surface defects be mistaken for bubbles?

Yes. Scratches, dust, and particles on the surface can produce similar image features. The standard verification technique is through-focus observation: internal bubbles shift laterally relative to the surface when the focal plane is moved, while surface artifacts remain fixed. Careful cleaning and re-inspection are used to resolve ambiguous indications.

What sample size and quantity are required for testing?

For coupon-based microscopic evaluation, samples of approximately 50 mm by 50 mm to 100 mm by 100 mm cut from defined sheet positions are customary. For lot release, either full sheets are inspected by AOI or a statistical sampling plan agreed with the customer is applied. Surfaces must be clean and free of handling damage before inspection.