BDI: 1,842 ▼ 1.2%
COTTON NO.2: 84.12 ▲ 0.4%
LME COPPER: 8,432.50 ▲ 2.1%
FOOD SAFETY INDEX: 94.2 ARCHIVE_SECURED
OPTICAL INDEX: 11,204.09 STABLE
BDI: 1,842 ▼ 1.2%
SECTOR INDEX
V.24.08 ARCHIVE
Optical lens surface roughness testing produces a meaningful Ra value only when the measured surface, spatial bandwidth, filtering, and evaluation length are defined. Two results expressed in nanometres are not necessarily comparable: a short scan across a polished patch and a longer scan that includes broader surface structure can produce different values on the same lens. Acceptance therefore depends on the complete measurement definition, not on Ra alone.
For a technical specification, identify the relevant optical surface and measurement region, the profile or area method, the processing applied to the data, and the required reporting conditions. Surface texture standards provide terminology and measurement frameworks, but a reference to a standard does not by itself establish a suitable roughness limit for every lens.
Ra is the arithmetic mean of the absolute height deviations of a roughness profile from its mean line over the evaluation length. In practical terms, it indicates the average magnitude of the profile’s departures from that line after the specified processing. It does not preserve the sign of those departures, and it does not describe their arrangement along the surface.
A profile with shallow, closely spaced polishing marks can have the same Ra as one containing a smaller number of deeper features. Their optical effects can differ because scatter depends on feature size, direction, surface material, wavelength, and illumination geometry. Ra also does not identify whether a prominent feature is a scratch, a pit, contamination, or a measurement artifact.
The distinction between Ra and Sa matters. Ra is calculated from a line profile; Sa is an arithmetic mean height parameter calculated over an area. Both are commonly reported in length units, but their different sampling geometries mean that their numerical values should not be substituted without an agreed method. A line can miss an isolated defect that appears in an area scan, while the area result depends on the field of view and data processing.

Contact and optical instruments each sample the surface through a measurement system with finite resolution and a defined response to surface slope. The choice should follow the lens geometry, material, expected texture, and need to preserve the finished surface.
A stylus trace requires particular care on a finished lens. A finite-radius tip cannot enter every narrow valley, so the recorded profile represents the surface as sampled by the tip. Excessive force or repeated tracing can damage a coating or alter a soft surface. A trace across a steeply curved region also places demands on instrument travel and slope capability. When contact is permissible, trace orientation should be recorded relative to visible polishing or machining marks.
Optical techniques avoid stylus contact, but a height map is still a reconstruction from an optical signal. Weak reflections, steep local slopes, transparent layers, and surfaces near the edge of an objective’s working range can produce invalid pixels or misleading heights. On a coated lens, reflections from different interfaces can affect the signal; the measured interface and any coating-related limitations should be established before comparing results with a substrate specification.
Atomic force microscopy offers a closer view of fine texture, yet its small scan area creates a sampling problem. A smooth patch does not establish that the full clear aperture has the same finish. Larger-area measurements and multiple selected locations are often needed when uniformity across the lens matters.
A lens is intentionally curved. The raw measurement therefore includes nominal shape as well as departures from that shape. Tilt and an appropriate form model are commonly removed before roughness is calculated. The choice of model matters: removing too little form leaves curvature in the residual profile, while removing an overly flexible fitted surface can suppress real long-wavelength structure.
Roughness filtering then separates the reported profile from longer-wavelength components according to the specified filter and cutoff. The cutoff is a spatial scale, not a universal dividing line between good and bad surface features. Changing it changes which features contribute to Ra. Short-wavelength response also matters: instrument resolution, stylus tip geometry, optical configuration, sampling interval, and any short-wavelength filtering determine how much fine texture enters the result.
For example, two measurements from the same polished glass region can disagree because one retains broader polishing structure that the other removes. That discrepancy does not automatically establish that either instrument is inaccurate. First compare the form removal, filter type and settings, sampling interval, trace length, and measurement location. Agreement in units alone is insufficient.
Edge handling deserves attention on small lenses. Filtering near the ends of a profile can be affected by limited data outside the evaluation region. Measurements close to a bevel or aperture boundary can also include edge artifacts or material outside the specified optical zone. The report should distinguish the recorded trace or image size from the portion actually used to calculate the parameter.
A single central measurement is convenient, but it may not represent a lens polished with a process that varies from centre to edge. Define positions within the clear aperture and record their distances or coordinates relative to the lens centre. For aspheric surfaces, include local slope and measurement accessibility when choosing locations. If different zones have different functional requirements, assign and report their results separately.
Direction is equally consequential. A line scan parallel to polishing marks can cross fewer height changes than a perpendicular scan. When texture has a preferred direction, use defined orientations or acquire an area map that reveals the pattern. Multiple traces also make it easier to distinguish a repeatable texture from a single local event.
Preparation should preserve the surface being evaluated. Loose particles, cleaning residue, fingerprints, and moisture can produce apparent peaks or interrupt an optical signal. Use a cleaning procedure compatible with the substrate and coating, and document it when the result is sensitive to preparation. If a feature disappears after permitted cleaning, the first measurement represented the surface condition at that time; it should not silently be treated as permanent substrate roughness.
Inspect invalid or missing data before accepting an area parameter. Interpolating across a small isolated gap and reconstructing a substantial unmeasured region are different operations. Report the extent and treatment of missing data, and avoid comparing a heavily reconstructed map with a complete map as though they sampled the same surface.
For smooth optical surfaces, roughness is often considered because surface irregularities scatter light. The relationship is governed by the height and spatial distribution of features as well as wavelength and measurement geometry. A low Ra does not guarantee low scatter at every angle, and an Ra limit chosen without the relevant spatial scales may fail to control the structure that matters in the optical system.
A power spectral density description can add detail by showing how surface variation is distributed across spatial frequencies. It is useful when two surfaces have similar Ra values but different polishing textures, or when a scattering model specifies a spatial-frequency range. The result still requires a stated measurement bandwidth and processing method; an instrument cannot describe frequencies it has not resolved or sampled.
Isolated scratches and digs require separate consideration. Averaging over a long profile or large area can dilute the contribution of a localized defect, while a trace that happens to cross it can produce a markedly higher value. Where local defects affect imaging, coating integrity, or scatter, evaluate them under an appropriate surface-imperfection specification rather than relying on Ra as their sole acceptance measure.
Likewise, surface roughness does not replace a form or wavefront measurement. A lens can have a smooth local finish and still depart from its intended shape. Conversely, a satisfactory form result does not establish that fine-scale texture meets a scatter-sensitive requirement. The relevant specifications should identify the scale and optical effect each measurement is intended to control.
A useful test report ties the numerical result to a reproducible measurement. It identifies the surface and coating condition, instrument and measurement mode, locations and orientations, profile length or area dimensions, sampling, form removal, filters, parameter units, and number of measurements. It also states whether the reported value is an individual result, an average, or the highest value among defined locations. An average can conceal a local failure if the acceptance rule applies to each location.
Measurement uncertainty becomes especially important when a result lies near a limit. Calibration or verification of the instrument, repeatability, environmental conditions, data processing, and surface variability all contribute to confidence in the comparison. The acceptance agreement should state how uncertainty is handled; a bare value rounded to a few digits does not resolve a borderline result.
When results from different instruments disagree, return to the original profiles or height maps where available. Compare measured regions, resolved spatial scales, filtering, and data quality before attributing the difference to a change in the lens. A shared test definition and an agreed acceptance rule make subsequent measurements more useful than an Ra value detached from its conditions.
Recommended for You