de gb
de gb

Non-destructive measurement of antireflective coatings on curved surfaces


By Lothar Völker, Michael Schulz-Grosser, Jos. Schneider Optische Werke GmbH, Bad Kreuznach & Rüdiger Kubitzek, AudioDev GmbH, Heinsberg

To minimize disruptive reflection losses, optical components such as lenses and filters are now equipped with highly effective antireflective coatings. However, the measurement methods used in quality control using scanning spectrophotometers are often non-destructive in practice and can hardly be used at arbitrary locations on a curved surface. A special measurement head design in conjunction with a diode array spectrometer provides a solution.

1. Classic Measurement Method
While coating technology has steadily advanced in recent years and quality control has become increasingly important, the methods for measuring the spectral reflectance of coated optics have remained largely unchanged. Reflection from the backside remains a significant problem during reflection measurements because it overlays the desired frontside reflection. For this reason, test samples are still coated in coating chambers and then roughened and blackened on the backside in time-consuming steps to suppress unwanted backside reflections1. These test samples can then be measured in a spectrophotometer, and the measurement values are assumed to be representative of an entire coating batch.

It is evident that with these destructive tests, optics to be used later cannot be tested for one hundred percent quality control. This becomes problematic especially when a customer complaint arises and there are no direct measurement data available for the questioned optic. Furthermore, even well-equipped spectrophotometers often measure only the central part of the substrate surface averaged over a relatively large area. Consequently, it can happen that the measured spectral reflectance curves deviate significantly from the actual curves outside the center of the lens.

Such deviation can occur, especially in the case of convex lenses with small radii of curvature, which are coated in a traditional bell jar without planetary gear in a coating system. Due to the different distances from the lens edge to the lens center and the deposition source, thinner coatings are generally deposited at the lens edge than at the lens center.

This azimuthal thickness distribution, which is approximately proportional to the cosine of the angle α between the surface normal and the deposition direction, results in a blue shift of the spectral reflectance curve of the antireflective coating. Hence, it is crucial, especially when using such lenses in objectives where the user desires minimal light loss across the entire field of view, to reliably know the actual course of the spectral reflectance not only in the lens center but also in the edge region of the lens.

1The necessity to roughen and blacken the test samples on the backside does not stem from the curved surface, but rather from the measurement at a 0° angle of incidence, where a spatial separation of front and backside reflection is fundamentally not possible. There are further reasons for the use of test samples, such as scratch resistance tests or adhesive tape tests to verify the adhesion strength of the coating.

2It is advisable to determine the reflectance at the lens edge for nearly perpendicular incidence in order to draw conclusions about the thickness of the coating there. In fact, in the case of lenses, light generally does not hit the lens surface perpendicularly at the edge but at partially shallow angles of incidence. This leads to an additional blue shift in the antireflective effect, which, however, is already taken into account by powerful optical design programs if the thickness variation towards the edge has been correctly inputted beforehand.

Lens system

Image 1: Spectral reflectance of a single layer of MgF2 on a commercially available plano-convex lens (diameter 50 mm, curvature radius ~50 mm) when measured at the lens center (red curve) and at the edge (blue curve). The spectrum from the lens edge is significantly blue-shifted.


2.Non-destructive Measurement

Many of the disadvantages of the classical destructive measurement method described above can be avoided by masking out the backside reflection through an innovative measurement head concept and replacing the scanning spectrophotometer with a diode array spectrometer. In such a device concept, broadband spectral light is guided to the measurement head via fiber optics and focused nearly perpendicularly, for example, at an angle of 8° to the incident slot, onto the lens surface (Image 3).

The light reflected from there under an angle of -8° to the incident slot according to the law of reflection is collected and guided via another fiber optic into a polychromator, where a diode array records the spectral intensity distribution in the exit slit plane. With focus and tilt sensors, the measurement head can be aligned at any point above the lens surface at the correct distance and angle, even outside the lens center2.

At the chosen angle of incidence of 8°, sufficient geometric separation of front and backside reflection is achieved. Only the frontside reflection is imaged onto the receiving optics. Moreover, the deviations from a measurement at 0° and the influence of the polarization direction of the incident light can still be neglected. An additional spatial filter helps in separating front and backside reflection, allowing even lenses with a thickness of only 2 mm to be measured well, without the need to roughen and blacken their backside (Image 4).

As a comparative measurement method, the achievable photometric measurement accuracy largely depends on the accuracy of the spectral reflectance of the used reference. It is recommended to choose a reference whose reflectance is within a few percentage points of the expected reflectance for the samples to be measured. Particularly suitable in this context are uncoated glass substrates of known materials. Its dispersion curve – and thus the spectral reflectance curve – can be calculated with very high accuracy using the Sellmeier coefficients tabulated in glass catalogs.

The measurement accuracy can be verified by referencing with an uncoated glass substrate of one material and then measuring an uncoated glass substrate of another material. A photometric accuracy of approximately ±0.5% of the spectral reflectance of the used reference is achieved. Therefore, if, for example, N-BK10 with R = 4% is chosen as a reference, an accuracy of R < ±0.02% is achieved.


3. Color Determination and Layer Thickness
Considering the spectrum ϕ(λ) = S(λ)·ρ(λ) captured by the spectrometer as the color stimulus function, where S(λ) represents the spectrum of the light source and ρ(λ) denotes the spectral reflectance, the normalized color value components x and y (the color locus), as well as the tristimulus values X, Y, and Z of the color stimulus of the coating in the CIE 1931 Standard Colorimetric System, can be calculated directly from this function, as shown in Image 5.

Determining the color locus during the measurement of antireflective coatings is indeed meaningful, as this allows for the selection of lenses from multiple measurements that possess a similar color locus. This is desirable, for example, in the assembly of binoculars or eyeglasses. With the presented measurement method, the thickness of a single layer can also be determined from the spectral reflectance curve, provided that the refractive index (or refractive index distribution) of the layer material is known. For layer thicknesses above approximately 1 μm, a Fourier transformation is used for this purpose, while for thinner layers, curve fitting is employed.

Furthermore, this option of determining layer thickness is also well suited for measuring the thickness of adhesive layers, provided that they are located between lens elements, each of which is at least 2 mm thick (Image 6). Since the diameter of the measuring spot is significantly smaller than 1 mm and the measurement method allows measurements at arbitrary points, it is even possible to investigate the homogeneity of the adhesive layer in this way.


4. Summary and Outlook
The presented innovative method enables very rapid, direct, and non-destructive measurement of the spectral reflectance at any point on curved, AR-coated lens surfaces through an intelligent measurement head concept.

The backside reflection of the sample is masked out. The ability to measure actual products rather than just specific test samples results in multiple additional benefits for quality control, development, and production. The measurement method achieves a similar high accuracy as conventional measurements with high-quality spectrophotometers without elaborate sample preparation, thus offering itself as an alternative or meaningful complement to the classical measurement method.

Typisches Messergebnis

Image 5: Typical measurement result showing the determination of the normalized color value components x and y as well as the tristimulus value Y (brightness) for a 2° standard observer and standard illuminant D65.


[1] M. Schulz-Grosser, Reale Transmission von Objektiven und optischen Systemen, Vortrag, 7. Arbeitstreffen des AK DUV/VUV-Optik, Dresden, 28. Sept. 2004
[2] H. Frey, G. Kienel (Hrsg.), Dünnschichttechnologie, VDI Verlag, Düsseldorf, 1987, S. 282
[3] J. Bliedtner, G. Gräfe, Optiktechnologie, Fachbuchverlag Leipzig, 2008, S. 49
[4] M. Richter, Einführung in die Farbmetrik, Walter de Gruyter, Berlin 1981, 2. Aufl ., S. 68 ff

Material float

Image 2: The material flow Φ of the evaporated materials through a unit area perpendicular to the connecting axis to the evaporating source is constant. However, the same amount of evaporated material is distributed on the lens axis and on the lens edge over areas of different sizes.



Prinzip des Messkopfs

Image 3: Schematic diagram of the measurement head, for simplification, focus and tilt sensors as well as the transmitted beam are not shown.



Optische Messung

Image 4: Non-contact optical measurement is successful even with relatively thin lenses, without the need to roughen and blacken their backside.




Image 6: Typical measurement curve for determining the thickness of an adhesive layer of an adhesive element. The reflected spectrum shows the interference-induced modulation typical for micrometer-scale layers.



Get in Touch

Please do not hesitate to contact us if you have any questions. Our dedicated team is here to help you every step of the way.
Whether you need assistance with product selection, technical specifications, or general inquiries.

Jos. Schneider Optische Werke GmbH
Ringstraße 132
55543 Bad Kreuznach | Germany

Tel: +49 (0) 671 601 205
Contact form