2-D + 3-D = Better Measurement

Stylus and optical surface metrology: the best of both worlds

Stylus profilers have been the primary tools for surface-finish measurement for many decades, providing cost-effective, rapid and quantifiable surface roughness and form information via single or multiple traces across samples. Most surface-finish standards are based on 2-D stylus parameters such as Ra, Rpm or Rz.

Figure 1: A DektakXT stylus profiler (left) and close-up of a stylus tip contacting a metal sample. All images courtesy Bruker Nano Surfaces.

Nonetheless, R&D and production operations have moved to 3-D optical surface metrology systems. This shift has been driven by growing demands on product function, lifetime, yield and appearance, which, in turn, require more comprehensive surface analyses. However, 3-D optical microscopy often complements rather than replaces contact methods. Both methods provide valuable information and combine well for comprehensive surface analysis, particularly in micromanufacturing.

This article discusses the primary attributes of each technique, the issues that must be addressed when introducing 3-D surface metrology and how 2-D and 3-D techniques can be combined for comprehensive surface characterization and process control.

Stylus profilometry

A stylus profiler is typically the first surface measurement instrument any organization purchases. The technology involves moving a small needle across a surface and noting its deflection as it travels. Very low contact forces, down to 0.03mg, are possible, so even soft samples are not damaged by the contact scan. Figure 1 shows a stylus system and a close-up of the tip contacting the sample. The instrument scans the sample underneath the stylus tip, in this case with motion towards the front of the instrument. Such systems are capable of sub-nanometer vertical resolution.

Lateral-stage motion enables multiple scans to be combined for areal sample coverage instead of simple traces. The stylus radius can be as small as 50nm, allowing high spatial resolution, but is typically about 10µm. Scan lengths for standard systems are up to 150mm (6"), but stitching multiple measurements together can be done to measure longer distances.

Figure 2: A 2.5nm step measured with multiple traces from a DektakXT stylus system to form a 3-D image.

Stylus profilers have several key advantages compared to optical systems. Because the measurement involves surface contact, effects from thin films, colored surfaces or material variations will not affect the result, unlike certain optical techniques. Stylus systems can even accurately measure thin-film steps down to 2.5nm in height (Figure 2).

Stylus systems are also ideal when only a single trace across a long distance is required, such as in an assessment of wafer or substrate bow and warp or other quick shape checks. Substrate measurements before and after deposition of a coating can even be combined to precisely quantify the sample stress from the coating process. In addition, stylus systems are often less expensive than optical systems, with entry-level systems costing about $30,000, compared to entry-level 3-D optical microscopes, which cost from $50,000 to $60,000.

3-D optical microscopy

While developed after stylus profilers, 3-D microscopes have been employed for surface characterization for more than 25 years. One of the most common forms of 3-D microscopy is based on coherence scanning interferometry, also known as white-light interferometry (WLI). Such systems use specialized microscope objective lenses that provide a signal related to the surface height as an objective is moved through focus.

Figure 3: A 3-D optical microscope and associated surface measurement of a titanium implant (left), corroded enamel (center) and microfluidic device (right) measured with a Bruker ContourGT 3-D optical microscope.

These systems provide sub-nanometer vertical resolution over fields of view from 50µm to more than 13mm. Measurement time for an entire field of view is typically under 5 seconds. Operation is simple: focus on the sample and adjust the light level as with any microscope, then tell the system to begin a measurement. It scans the objective through focus and produces a 3-D map of the surface.

Due to the maturity of the technology, these systems have comprehensive analysis packages to meet almost any application need. WLI systems provide the most-accurate data over the largest field of view and the fastest time-to-result of any 3-D metrology system.

Surfaces with steps up to 10mm, such as some stepped machined parts in the automotive industry, can be readily measured with a WLI system. They can handle most any surface roughness as well, measuring everything from extremely smooth mirrors to objects as rough and dark as black-vinyl foam. Large fields of view can be measured via stitching; even 300mm-dia. wafers have been fully mapped with this technique. Figure 3 shows representative measurements on a variety of surfaces.

Adoption of 3-D microscopy

Three-dimensional optical microscopes often provide measurement capability that cannot be readily provided by contact methods. In some cases, such as measuring solar cell efficiency, 3-D “S parameters”—the ISO-standard 3-D extensions of the original 2-D “R parameters”—used with stylus instruments were shown to directly correlate to solar cell efficiency, while traditional methods, such as counting key features or calculating Ra, did not. Three-dimensional optical microscopes can also measure down steep slopes—those greater than 70°. This allows the angles between critical interfaces in tooling or other parts with steep features to be measured.

Scratches, defects, wear scars or corrosion sites can be measured, and both the dimensions and the volume of those features can be determined quickly and precisely via 3-D microscopy. Such systems even allow measurement of surfaces through glass or other transparent media.

Figure 4a: Stylus (upper left) and 3-D microscope (upper right) images of a polysilicon solar cell. Figure 4b (bottom) shows the close agreement of 3-D average roughness (Sa) for stylus and optical measurements.

A last critical advantage of 3-D systems is that the roughness value obtained is independent of part orientation, with respect to the measurement head; with a 2-D stylus scan, the operator must precisely align the scan direction perpendicular to critical features on the surface so as to not distort the results.

Correlation between 3-D optical microscopes and 2-D stylus systems has been established through a variety of studies. Each measurement must be carefully controlled and filtered so the differences in lateral resolutions between the systems are matched.

However, this process is relatively straightforward. Figure 4a shows measurements of a solar cell taken with a Bruker DektakXT stylus system and ContourGT 3-D optical microscope. The results from the two techniques closely agree. Also, across a variety of surfaces with different finish processes, surface roughness (Sa) correlates well (Figure 4b).

The value of any new parameters must also be clear before any existing specifications are replaced or additional ones are introduced. The process is typically reactive rather than proactive; only when failures occur (in terms of lifetime, emissions or appearance) are new parameters identified to better control the process.

A common procedure when an organization acquires a 3-D microscope system is to log all 2-D parameters, in order to maintain continuity with prior methods, as well as logging additional 3-D parameters identified jointly with the metrology equipment manufacturer as relevant to that application. When failures to overall quality goals are discovered, the measurement history is analyzed and often it is seen that one or more 3-D S-parameters correlate closely with good or bad part identification.

In other applications, the additional information provided by fast, accurate, large-field surface data is able to quantify processes that could not be controlled via 2-D metrology. In the solar industry, for example, only by use of 3-D S-parameters could surface roughness finally be quantified in terms of correlation of roughness to solar cell photovoltaic efficiency (Figure 5).

Figure 5: An etched solar cell measured at low-magnification and correlation of surface skewness (Ssk) to efficiency.

When measuring cylinders used in dynamic seals, the “lead angle” is measured. (This angle is formed by the cylinder and the surface structure produced by the cutting or grinding process.) The traditional way to quantify the lead angle is to wrap a weighted string around the cylinder and watch it unwind as the cylinder rotates. Three-dimensional optical profilers, by comparison, are able to quantify micro-lead values, with the added benefits of being self-referencing and measuring critical surface-roughness parameters simultaneously.

Other industrial applications with tight tolerances include aerospace components, medical implants, MEMS devices and ophthalmic devices. Manufacturers of these parts and devices are rapidly adopting 3-D optical microscopy.

Stylus profilometry is a well-established, cost-effective technique for surface measurement. Despite persistent speculation in the metrology world that 3-D optical microscopy will replace stylus profilers, they are still the dominant surface-measurement tools in terms of numbers of units employed. And they continue to perform extremely well in many applications.

However, 3-D optical microscopes based on white-light interferometry often provide critical throughput enhancements and surface analyses that cannot be obtained from stylus traces. The methods can be shown to correlate, and the maturity of the 3-D optical techniques makes them easy to use. This allows them to provide consistent, operator-independent results.

With ISO and other standards groups supporting measurement via both techniques, I expect 3-D measurement to continue to complement stylus techniques in critical industrial and research applications. µ