Optical Glass Striae Detection: Testing Methods, Standards, and Quality Grading

Optical glass striae detection is a critical quality control process used to identify thread-like refractive index inhomogeneities that degrade wavefront quality, image resolution, and laser beam integrity in precision optical systems. Using shadowgraph projection, Schlieren, and interferometric techniques in accordance with GB/T, ISO, and MIL-G-174 standards, accredited third-party laboratories and manufacturers grade striae severity in raw glass blanks and finished optical components. This article explains the formation mechanisms of striae, sample preparation, testing items, detection methods, and typical industrial applications.

Technical Overview

Striae (also called veins or cords) are localized, thread-like or streak-like regions within optical glass where the refractive index deviates slightly from the surrounding matrix. The refractive index difference (Δn) inside a stria typically falls in the range of 10⁻⁵ to 10⁻³, while the physical width may vary from tens of micrometers to several millimeters. Because striae act as weak embedded lenses, they deflect and distort transmitted light and are among the most damaging bulk defects in high-performance optical systems.

Striae originate primarily from the glass manufacturing process. Common causes include incomplete homogenization of the melt due to insufficient stirring, dissolution of crucible and refractory materials into the glass, volatilization and re-condensation of easily evaporated components such as alkali and boron oxides at the melt surface, thermal gradients during forming, flow lines generated in pressing or molding operations, and non-uniform fine annealing. Once the glass has solidified, striae cannot be removed by any subsequent heat treatment.

Functionally, striae introduce wavefront distortion, reduce modulation transfer function (MTF), generate ghost images and local scattering, and in laser applications degrade beam quality and propagation characteristics. Their effect is amplified in thick components such as prisms, windows, and waveguide substrates where the optical path length through the defect is long.

Striae detection is therefore a standard element of optical glass acceptance inspection. Relevant reference frameworks include the GB/T 903 specification for optical glass and the GB/T 7962 series of test methods for colorless optical glass (which defines the shadow-type striae test), ISO 12123 covering optical glass vocabulary and related attributes, and the striae quality grades A through D historically defined in MIL-G-174. Individual glass manufacturers further assign catalog striae classes based on comparison testing against certified reference samples.

Testing Samples

Raw Glass Blanks and Pressings

Blocks, strips, plates, rolled sheets, and press-molded gobs are tested in the as-delivered state before cutting or grinding. Testing raw stock is the most economical approach, as bulk striae quality can be verified before value is added in machining. Typical sample thickness ranges from 20 mm to 100 mm, with thicker samples providing a more conservative assessment because striae contrast grows with path length.

Semi-Finished and Finished Optical Components

Lenses, prisms, plane-parallel windows, filters, reticles, and beamsplitter substrates can be tested after grinding or polishing. Components are oriented so the probing beam traverses the longest practical optical path. Curved surfaces may require compensation optics or index-matching immersion to separate surface effects from internal striae signatures.

Specialty Optical Materials

Materials such as fused silica, high-index lanthanum crowns, dense flints, phosphate and fluorophosphate laser glasses, and low-transformation-temperature glasses for precision molding are frequently submitted for striae evaluation. Glasses containing volatile components are inherently more prone to striae formation and are commonly subject to tighter inspection regimes.

Sample Preparation Requirements

• Two opposite faces should be parallel and either polished or fine ground to a defined surface roughness.
• Fine-ground surfaces are tested together with an index-matching liquid to suppress surface scatter and reveal internal structure.
• All four side faces are ground opaque and chamfered; edges are cleaned to avoid false indications.
• Samples are tested in at least two, and preferably three, mutually perpendicular directions, since striae are often oriented in layers parallel to the forming direction.
• Excessive residual stress should be relieved by appropriate annealing before measurement, and samples must be free of oil, fingerprints, and moisture films.

Testing Items

Depending on the specified inspection level, a striae detection program may include the following evaluation items:

• Striae class or grade assigned by comparison with certified reference standards under defined test sensitivity.
• Dimensions of individual striae, including width, length, and orientation relative to the optical axis or datum faces.
• Striae density, expressed as the count of visible striae per unit cross-section or per sample face.
• Striae contrast or ray deflection capability, reflecting the local refractive index gradient.
• Optical path difference (OPD) introduced per unit thickness and the derived local Δn, obtained by interferometric methods.
• Accompanying stress birefringence, since striae are frequently associated with localized stress in the glass.
• Conformity judgment against the striae tolerance stated on the optical drawing or purchase specification.

Testing Methods

Shadowgraph (Parallel-Beam Projection) Method

The shadowgraph method is the classical and most widely used striae test. A point or slit light source is collimated into a parallel beam that passes through the sample before falling on a viewing screen placed at a defined distance. Striae deflect rays out of the main beam, producing characteristic dark or bright thread-like shadows on the screen. Sensitivity is adjusted by varying the source size and the sample-to-screen distance, and the observed shadow width and contrast are compared with graded reference samples to assign the striae class. The method provides a full-aperture view, is fast, and forms the basis of the striae grading defined in the GB/T 7962 series and MIL-G-174 style specifications.

Schlieren (Knife-Edge) Method

In the Schlieren configuration, the transmitted beam is focused onto a knife-edge or cutoff stop placed at the focal plane of the imaging optics. Small angular ray deflections caused by striae are converted into measurable intensity variations in the image plane. Because the cutoff rejects the undisturbed background light, the Schlieren method achieves higher sensitivity than simple shadowgraphy and allows mapping of the refractive index gradient distribution. It is preferred for tighter striae classes and for research into homogenization processes.

Interferometric Methods

Twyman-Green, Mach-Zehnder, and lateral shearing interferometers measure the optical path difference introduced by striae in transmission. With phase-shifting fringe analysis, the local Δn can be calculated directly from Δn = OPD / t, where t is the sample thickness. Interferometry provides fully quantitative results with detection capability extending to approximately ±1 × 10⁻⁶ in refractive index under controlled conditions, making it suitable for high-homogeneity glass grades, thick prisms, and lithography-class optics. It simultaneously yields the overall homogeneity class of the sample.

Immersion and Visual Comparison Techniques

Small components, ground blanks, or irregular pieces may be immersed in an index-matching liquid and observed under directional dark-field or side illumination. Matching the immersion liquid to the glass suppresses surface refraction so that internal striae appear as visible contrast lines. Side illumination perpendicular to the viewing direction is particularly effective for revealing fine layered striae. This technique is also used for rapid incoming inspection and for verifying reference samples.

Supplementary Stress Birefringence Measurement

Because chemical inhomogeneity and stress frequently coexist, a crossed-polarizer or photoelastic polarimeter measurement expressed in nm/cm is often performed alongside striae testing. Elevated stress readings localized at striae positions help confirm defect severity and indicate whether additional annealing or melt process correction is required.

Applications

Striae detection supports quality assurance across the entire optical manufacturing chain. Major application areas include:

• Imaging optics: camera modules, telephoto and zoom lenses, binoculars, telescope objectives, and microscope objectives, where striae directly reduce resolution and contrast.
• Projection and AR/VR systems: pancake lenses, prisms, and waveguide substrates, in which striae produce visible streaks, ghosting, and non-uniformity across the eyebox.
• Laser and photonics: resonator optics, Q-switch windows, harmonic generation crystals housings, beam delivery windows, and high-power laser components where wavefront purity is critical.
• Aerospace and defense: reconnaissance lenses, FLIR windows, periscope and sighting prisms, and head-up display optics requiring stringent bulk homogeneity.
• Semiconductor and metrology equipment: illumination optics, interferometer reference components, and stage windows used in lithography and precision measurement instruments.
• Automotive and medical: LiDAR scanner windows, ADAS camera optics, endoscope rod lenses, and surgical microscope objectives.

Glass manufacturers rely on routine striae testing to validate melting, stirring, and annealing parameters, while lens fabricators and system integrators use it as an incoming inspection gate to prevent defective glass from entering production.

FAQ

What are striae and how do they differ from bubbles and inclusions?

Striae are continuous, thread-like variations in refractive index caused by chemical or thermal inhomogeneity in the glass melt. Bubbles and inclusions are discrete, geometrically bounded defects such as gas voids, stones, or crystalline particles. Striae scatter and deflect light gradually along their length, whereas bubbles and inclusions scatter light at their boundaries. The two defect types are graded by separate test methods and separate tolerance classes on optical drawings.

How is striae quality expressed?

Striae quality is expressed as a class or grade determined by comparison testing. Class definitions combine the shadow visibility of the defect with the sensitivity of the test setup, typically controlled through source geometry and screen distance. Higher grades correspond to stricter requirements, with the best grade requiring that no striae be detectable at maximum test sensitivity. MIL-G-174 historically defined grades A through D, while Chinese national standards and manufacturer catalogs define comparable class scales.

Do sample surfaces need to be polished?

Polished parallel faces give the cleanest results, but fine-ground faces are acceptable when tested with an index-matching liquid applied between the sample and a polished cover plate or in an immersion bath. What matters is that surface scatter is suppressed so the shadow pattern reflects internal striae only.

Why must samples be tested in several directions?

Striae commonly form as layers parallel to the flow direction during casting, pressing, or rolling. A beam traveling parallel to the striae layers may miss them entirely, whereas a beam traveling perpendicular to the layers crosses many striae and reveals the true defect density. Testing in three mutually perpendicular directions ensures that no orientation of striae escapes detection.

What level of striae quality is required for demanding applications?

General photographic and instrument optics usually tolerate a moderate striae class verified at standard sensitivity. High-end applications such as lithography optics, interferometer components, high-power laser systems, and AR waveguides require the highest grade, in which no striae are observable at maximum shadow or Schlieren sensitivity, and often demand interferometric confirmation of homogeneity on the order of 10⁻⁶ in Δn. The required class should always be specified on the optical drawing in accordance with the applicable ISO 10110 indication practice or the purchasing specification.

Can striae be tested in finished lenses with curved surfaces?

Yes. Curved components can be evaluated by immersing them in a matched-index liquid, by using compensating optics that cancel the lens power, or by transmitted-wavefront interferometry, which integrates striae effects into the measured wavefront map. Interpretation must account for the longer effective path through steep surface regions.