Performing 3D Optical Profiling Surface Measurements that Correlate to Traceable Standards

This article explains the benefits of the non-contact inspection method used by 3D optical profilers, and outlines the best practices and measurement results for some specialized PTB (Physikalish-Technische Bundesanstalt) traceable roughness standards and other low-cost fingernail roughness gages.

The correlation results are based on measurement factors that should be understood and taken into consideration when imaging and analyzing surface textures that range in roughness from a few nanometers to micrometers in scale.

3D optical profilers that utilize coherence scanning interferometry, also called white light interferometry (WLI), deliver fast, accurate surface measurements over large areas to determine various properties about surfaces of interest. They are being increasingly employed in research, engineering, and production process control for a variety of markets, including materials science, aerospace, automotive, data-storage, solar, semiconductor, and MEMS.

Understanding how this technology correlates to conventional 2D techniques and standards, and how the increase of measurement data can be determined and utilized is vital to fully exploit the capabilities of today’s top-performing 3D optical profilers.

Advantages of 3D Optical Profiling over Other Measurement Technologies

2D stylus profilers were used initially for surface roughness characterization in the early 1930s and were adopted as the industry standard until the development of 3D metrology instruments decades later. 3D optical profiler measurement systems have many advantages, which have prompted the international metrology community to develop new measurement standards to fully exploit this superior technology.

Modern sophisticated surface profilers have industry-leading speed and accuracy, while maintaining the same “nanometer” Z accuracy at all magnifications. Such systems are capable of measuring a very wide range of surface parameters, including pitch, roughness, curvature, step heights, waviness, and lateral displacement, all in a single measurement and on almost any surface.

Based on white light interferometry shown in Figure 1, this measurement technique is capable of rapidly determining 3D surface shape over large lateral areas - up to 8 mm - in a single measurement. Surface areas larger than this value can be measured by applying stitching algorithms to enable multiple lateral images to be captured and merged into one image for analysis.

Although several other technologies provide fast speeds, good resolution, or larger areas of measurement, they each have their own limitations as well. For instance, stylus profiling provides scans up to hundreds of millimeters in length, but each scan is just a trace along one probe-tip-wide line, limiting the area that can be analyzed without taking multiple traces, which makes acquisition time slow for larger areas.

Figure 1. Basic white light interferometry schematic with Bruker’s self-calibration HeNe laser.

Similarly, confocal microscopes provide reasonable Z resolution at very high magnifications, but have much slower data acquisition time because of the scanning technology used to capture the Z height data. Finally, optical focusing techniques are employed for more coarse manufacturing surface finishes, but are not generally able to achieve the Z resolution of an interference-based 3D optical profiler, especially for surface texture on finely machined structures.

These other techniques also have other downsides when it comes to measuring surface topography and quantifying texture. A major disadvantage of contact stylus measurement is that the stylus tip has to run perpendicular to the predominant surface pattern or surface lay of the measurement surface. If this is not the case, the tip may follow the surface structure and deliver false surface texture results, similar to a record player needle following the grooves in a record.

Another disadvantage of stylus measurements is the limitation in Z height measurement range. A stylus system needs to use a skid plate to extend the measurement range, allowing it to measure over larger steps but then limiting its ability to measure waviness or stepped features precisely as the skid plate must track the surface of interest. This produces a sort of mechanical filtering of the surface representation.

Finally, as very hard materials, such as diamond, are used to make most stylus tips in order to reduce wear and increase tip life, carrying out scans using them can cause damage to the surface of interest and deliver false readings, as shown in Figure 2.

Figure 2. Stylus damage to reference standard.

Confocal microscopy finds the height at each pixel location by calculating the center of mass of the intensity distribution around the focus position or by detecting the peak intensity. The intensity envelope is very narrow for high-magnification objectives, but becomes wider at lower magnifications due to the lower numerical aperture (NA) of the objective, which increases the depth of field.

This large depth of field impairs a confocal system's ability to repeatedly detect the centroid and peak intensity, and as a result, deteriorates the Z accuracy and resolution. High-magnification objectives (20x and above) must typically be used to gain Z accuracy, but this limits the lateral field of view.

Previously, the ability to measure steep angles was the key advantage of using a confocal microscope. However, the development of higher magnification objectives for 3D optical profilers as well as the improved lateral resolution of high-resolution cameras enables non-confocal systems to measure steep angles approaching 90° on non-mirror surfaces.

Modern 3D optical profilers do not have any limitations to surface structure orientation and eliminate surface damage, as they are based fully on a non-contact measurement technique. Additionally, 3D optical profilers are generally not Z-height limited, and are capable of measuring up to 10 mm in height.

The fringe envelope of a 3D optical profiler remains very narrow at all magnifications, from 0.75x to 230x, and as a result, the profiler maintains the same high Z resolution across the field of view at any magnification.

Stylus Profiler Filtering for Traditional Measurement Standards

Most of the traditional measurement standards have been set around contact stylus results. Understanding the working principle of this technology and the origin of the standards that came from it is needed in any effort to correlate other technology results.

A stylus contact measurement system possesses natural mechanical filtering due to the tip radius and taper angle of the stylus tip making contact with the surface during measurement. The taper angle is typically 45° to meet recommended measurement standards, such as International Organization of Standardization (ISO).

The tip radius range varies between 1 and 10 µm. Based on the sample roughness and tip model, the tip may not reach to the bottom of the surface profile, and can also round off the peaks and valleys, both of which will have an impact on the surface finish results. Also, measuring surface features that are smaller than the tip itself is not possible, as illustrated in Figure 3A and 3B.

Figure 3A. This 25 µm tip easily measures larger trenches (a), but cannot accurately measure the width (b) and height (c) as the trench aspect ratio increases.

Figure 3B. Influence of tip radius with stylus measurements.

It is necessary to determine a known scan length when setting up a mechanical stylus measurement. This scan length is the length of the path traced by the stylus and is known as the measurement or traversing length. The sampling length is usually defined as the spatial wavelength of the lowest spatial frequency filter that will be used for data analysis.

According to most industry-recommended practices or standards, the measurement length must be at least seven times higher than the sampling length or the wavelength of the feature of interest. As can be seen in Figure 4, one sample length is commonly discarded from each end of the measurement length.

Figure 4. Total profile with divisions into sample, evaluation, and traversing lengths.

The scan captured is known as the total profile and is electronically leveled, usually by fitting a line through all of the data. Applying electronic filters and cutoff filters is the next step in the analysis of stylus data.

Generally, the first step of analysis is to apply a low pass spatial frequency filter to the raw total profile to remove very high spatial frequency data, as it can often be attributed to stylus deformation, vibration, or debris on the surface. The next step is separating the data into waviness, roughness, and form by using various other filters.

Typically, a high-pass spatial filter is applied to remove the overall form or waviness to obtain the roughness parameter. The waviness profile can be obtained using a band-pass filter. Although many other types of filters are available, this is the standard approach when carrying out contact stylus surface finish measurements and analysis, as shown in Figure 5.

Figure 5. Electronics filters are applied to obtain roughness or form.

Figure 6A and 6B illustrate the most common types of manufacturing processes that have a similar surface finish result. Recommended practices outlined by standards, such as ISO and ANSI/ASME, suggest sample lengths, filters, and cutoffs for those stylus measurements.

When trying to correlate a non-contact measurement to a stylus measurement system, users have to know exactly how the stylus system was configured electronically and mechanically when collecting the data to offer the best reproducibility and correlation.

Figure 6A. Surface finish tolerances in manufacturing.

Figure 6B. Vertical milled surface measured on a 3D optical profiler.

For instance, the Vision64^® software of Bruker’s 3D optical profilers can be configured to replicate a 2D stylus measurement by applying those same filters and cutoffs, as shown in Figure 7. Due to the developments in optical profiling over the last decade, 3D optical profilers now have the capability to accurately perform 2D parameter measurements traditionally done by stylus profilers, with significant improvements in terms of speed of measurement and sample integrity.

Figure 7. Stylus analysis settings in Vision64 software.

Utilizing Optical Profiling to Extend Surface Parameter Characterization

With the long history of 2D stylus surface measurements providing such basic parameters as roughness (Ra), these statistical parameters typically end up being applied to measure and control the quality of surface finishes in many markets. Now, the development of sophisticated 3D surface characterization enables accurate measurement of those 2D parameters and also, provides additional data to advance the way a surface gets characterized.

These new 3D parameters can highlight large data trends, such as lay or waviness, and additional features, such as a predominance of ridges and scratches, that 2D traces are unable to characterize. New surface parameters are now increasingly used due to these 3D capabilities.

Known as “S Parameters,” these generally categorize hybrid, spatial, amplitude, and functional parameters (Figure 8). 3D parameters uniquely discern not only shape and surface finish, but also the functionality of that surface.

Figure 8. Typical applications for 3D parameters.

Correlating to Fingernail Surface Finish Gages

Recommended measurement practices for 3D optical profiling are being approved and adopted for many industrial markets. The capability of correlating to known traceable standards is required for most industries striving to achieve traceability and certification.

A fairly low-cost approach many companies apply is purchasing a surface roughness standard patch, which can come with a traceable certificate of calibration. These patches have Ra surfaces ranging between 50 nm to 13 µm (2 to 100 microinches), and different machined surfaces (Figure 9).

Figure 9. Multi-patch fingernail roughness comparator standard.

Theoretically, this seems to be a good low-cost approach for traceability and correlation, but it can lead to issues because these types of gages are meant to be a “fingernail” comparison standard for an operator to roughly compare the actual surface being machined.

An operator or engineer will actually compare the machined surface to the roughness patch using their fingernail or simply by the visual appearance. When an independent stylus gage calibration was performed on one of these patches using a recommended internationally approved measurement practice, a significant variation between the new traceable certification values and the original certification that came with the patch was observed, as shown in Figure 10.

No clear documentation is available that shows how the original certification values were obtained, but it appears that standardized measurement practices were not used to certify the gage.

Figure 10. Examples of original versus independent certification.

Once the actual values of the roughness patch are known, the measurement setup parameters applied during calibration must also be known so that they can be applied to duplicate the values on any system, in this case a 3D optical profiler.

When to Use Single Field of View Versus Larger Stitched Image Measurements

It is possible to measure lower roughness surfaces below 254 nm (10 microinches) with a single image on a 3D optical profiler because there typically isn't a manufacturing periodic machining structure or lay to worry about. Also, lower Ra surfaces tend to correlate well to contact stylus measurements even without applying the filters used by the stylus system.

One thing that has to be taken into consideration is that by going to higher objective magnifications, the lateral resolution pixel size gets smaller, allowing more of the microstructure of the surface to be seen. Potentially, as mentioned before, this microstructure cannot be measured by the contact system. Therefore, the system could reduce the correlation to the optical system.

To measure surface finishes with higher Ra values that may possess a periodic machining structure, as seen in Figure 11, the measurement area needs to be large enough to capture this periodic structure. This would be the same as a contact stylus profiler’s sampling length to meet standardized practices, as mentioned before.

This can be easily performed using a lower magnification objective to capture this entire periodic structure. Here, the tradeoff is the lower the magnification of the objective, the lower the objective NA, which, in turn, lowers the angle at which the objective can capture light and the angle at which it can measure slopes (similar to large radius contact probe tip limitations).

Once the Ra of the surface gets high enough, the structure of the surface gets very steep and the lower magnification objective may have data dropout similar to the stylus tip radius rounding or smoothing of the surface edges. This can be resolved by “stitching” multiple images together using higher magnification to cover the entire periodic structure area, similar to a stylus system sampling length.

Figure 11. 10x stitched image of a turned surface.

Proper Filtering to Correlate to Contact Measurement

Even after an independent certification on the fingernail roughness patch is performed, it may not be enough to merely compare the 3D optical profiler roughness directly to that certification. However, the filtered 3D data can be well correlated to the certified stylus measurements once the proper scan length data is captured and the corresponding stylus filters are applied in the 3D optical profiler software.

This can be seen in Figure 12A, which depicts RA data before and after the stylus filter is applied, and in Figure 12B, after the filters were applied and correlated to the independent certification. The correlation results show a slight deviation that is attributed to not knowing exactly where the independent certification was conducted.

Figure 12A. Optical stylus measurement applied versus raw optical measurement.

Figure 12B. Optical stylus measurement versus independent certification.

Correlating Measurements to Nationally Traceable Standards

With regard to quality assurance, traceability of surface finish measurement plays a key role in manufacturing for the correlation and functionality of many products and parts. The ability to correlate surface finish parameters across the globe is carried out using standards that are traceable to nationally recognized bodies, such as Physikalisch-Technische Bundesanstalt (PTB) and the National Institute of Standards and Technology (NIST).

For this analysis, high-precision standards were bought and certified at PTB (Figure 13). The standards are produced for specific surface roughness, where the waviness and texture are highly repeatable and controlled across the measurement surface.

PTB provided calibration paperwork with detailed information on how the standards were certified using a contact stylus system. This certification data is very useful when trying to reproduce and correlate to these standards using various measurement techniques, including 3D optical profiling based on interference technology.

For all of the standards determined in the study, Bruker’s Contour Elite^® 3D optical profiler was used with a 10x objective and vertical scanning interferometry (VSI) mode. The data was captured as per the PTB certification, ensuring that it was captured in the proper location using the right measurement scan lengths (stitched images) and stylus filtering applied to the measurements.

After scanning the nine measurement locations on each standard, the results were averaged and compared with the error from the PTB certification uncertainty. The results of which are shown in Figure 14.

Figure 13. Precision surface roughness standard and measurement locations.

Figure 14. WLI deviation results as compared to certified roughness standards.

As observed in the deviation chart, correlation to nationally accepted roughness standards can be easily obtained provided that the proper measurement methods and certification process of the standard have been followed for the 3D optical profiler-based measurement.

Conclusion

Recent developments in 3D measurement techniques have provided researchers, engineers, quality control professionals, and process designers with an improved way to characterize surfaces for surface finish, shape, and overall functionality. 3D optical profilers are well established across a wide range of industries, ranging from aerospace components to medical implants, and have been shown to perform better than other measurement methods in terms of the overall speed, accuracy, resolution, and repeatability.

Up front knowledge of the surfaces being measured and the setup of the stylus tool used to measure those surfaces can help to achieve correlation to stylus measurement systems. Due to texture coherence/speckle effects and measurement location, some minor correlation differences are expected. However, these can be accounted for or tracked through a correlation factor, if required.

The addition of the 3D surface S parameters significantly extends the degree to which surface analysis can characterize both shape and function of a sample. The outcome is the most radical improvement in measurement data since the early introduction of 2D stylus-based profilometry.

References

1. ISO 4288, Geometrical Product Specification (GPS) – Surface texture: Profile method – Rules and procedures for the assessment of surface texture.

2. Surface Texture (Surface Roughness, Waviness, and Lay), American National Standard, ASME B46.1-2002, New York, NY.

3. ISO 4287, Geometrical Product Specification (GPS) – Surface texture: Profile method – Terms, definitions and surface texture parameters, 1997.

4. S. Bui and M. Novak, “ISO-Standardized Filtering for Dektak Stylus Profilometers,” Bruker Application Note AN550, 2014.

5. J. Petzing, J. Coupland, and R. Leach, “The Measurement of Rough Surface Topography Using Coherence Scanning Interferometry,” National Physical Laboratory, 2014.

This information has been sourced, reviewed and adapted from materials provided by Bruker Nano Surfaces.

For more information on this source, please visit Bruker Nano Surfaces.