Optical microscopy measures up as life-science devices miniaturize

The miniaturization of life-science devices is following a technology trend similar to that of the semiconductor and microelectronic industries. Devices come in various forms such as catheters, balloons, wires, cables, and sheets. Often these devices require drilling of micron-sized holes, narrow kerf (cut) widths, and small, shallow etched areas.

Molecular diagnostic products such as continuous glucose monitoring sensors, DNA sequencing, and polymerase chain reaction (PCR) platforms also require micron-sized features in order to isolate, capture, and detect the analyte. Laser micro-drilling is one of the few practical and scalable methods to produce tight-tolerance holes as small as 1µm in diameter.

Laser drilling uses photo-chemical or photo-thermal ablation where the material is vaporized into a plasma plume and expelled into the surrounding space. Using lasers with wavelengths in the ultraviolet region and short temporal pulse durations permits machining of micron-sized features without collateral heat effects in the surrounding material, an essential requirement to prevent distortion of the machined feature.

Life science products must go through a rigorous validation process. Confirmation by examination and provision of objective evidence are essential steps in process and design development and verification. To validate the laser drilling process and demonstrate its capability, accurate measurement techniques must be employed to demonstrate compliance to customer specifications.

Accurate measurement of holes in the micron range requires analysis of an image rather than conventional approaches such as contact coordinate measurement systems, pin gages, or calipers that are generally useful for measuring larger holes. Image analysis techniques include optical microscopy and scanning electron microscopy (SEM).

Optical microscopy has several advantages in comparison to SEM; most of these advantages relate to the relative cost of matching various optical capabilities with SEM, and this discussion will primarily reference lower-cost SEM systems. Optical systems are much easier to place within in-line production, allowing real-time measurement during the production process. Even if the inspection is offline, optical inspection is still typically faster. SEM inspection is performed in a vacuum and may require sputter-coating of non-conductive materials such as polymers with a thin metallic overlayer, which is destructive to the sample. Optical inspection, however, can nearly always be non-destructive, with witness or gold-standard samples to help set manufacturing guidelines.

SEMs, on the other hand, have one critical advantage. The resolution of any microscope is related to the wavelength of the light (or electrons) used to inspect objects. In the case of optical microscopes, the best results are obtained using monochromatic light, usually green, at approximately 550nm. With a reasonable optical setup, this wavelength results in a lower limit on optical resolution of approximately 0.20µm. Using the standard rule for inspection measurement of a factor of ten between resolution and measurement, a 2µm hole diameter would be the lowest measurement limit. For non-planar objects, resolution is affected by image defocusing, and across-the-field distortions may affect the measurements. Using automated, repeatable methods of size measurements (contrast-based edge detection, averaging between multiple measurements for the same feature) can improve measurement quality.

By contrast, a 10kV electron in a SEM has a wavelength of approximately 12 picometer, permitting measurements down to less than a nanometer using the same rule of thumb. For this reason, holes in the range of a few micrometers in diameter or less have typically been measured using SEM, since the measurement resolution is less than 1/1,000th the measured diameter, eliminating any possibility of resolution-related error. SEM, however, is not free of its own sources of error. For example, scale variation from changes in observation angle or working distance can introduce distortion into the image analysis. Given the high resolution available with SEM, even variation in a hole’s edge definition may contribute to the quality of the measurement.

Some of the measurement uncertainties in both optical and SEM microscopy can be mitigated by using on- or near-the-target reference features such as standard single or periodic arrays.

For this study, the following materials were laser-drilled:

i) Polyimide film, 0.005" (12.5µm) thick, target of 4.5µm diameter holes (Figure 1)

ii) Glass, 0.15mm thick, target of 2µm diameter holes (Figure 2).

All holes were measured at the hole exit side.

An automated optical CNC video measuring system was used to evaluate nine holes, laser-drilled through a 0.005mm-thick sheet of Kapton film; each hole had a nominal 4.5µm diameter. Each hole was measured 10 times by the automated optical system based on image contrast and edge detection. For comparison, all nine of the laser-drilled holes were also measured using a SEM to obtain nominal reference diameter values. A similar evaluation was performed on the glass sample with nominal 2.0µm laser-drilled holes, with 10 repeat measurement runs and SEM verification measurements.

The 10 repeat measurements performed by the automated optical system are indicated by the black data points in the plot above (Figure 3). The nominal diameter values established by SEM measurement are represented by the red reference lines.

The measurement repeatability of the optical system for the 4.5µm holes was evaluated using the Gage R&R ANOVA method, the results of which are summarized in the tables below and in Figure 4 above.

Shown in the Paired T-Test and CI: SEM, CNC Video Measuring System table below, and in Figure 5 below, is a statistical comparison of the optical vs. SEM hole diameters using a paired t-test.
 

The 10 repeat measurements performed by the automated optical system are indicated by the black data points in Figure 6 (right). The nominal diameter values established by SEM measurement are represented by the red reference lines.

The measurement repeatability of the optical system for the 2.0µm holes was evaluated using the Gage R&R ANOVA method, the results of which are summarized in the tables and Figure 7 below.

The limited correlation between SEM and optical measurements supports the claim that optical methods can accurately and repeatedly measure geometries down to 2.0µm with tolerances of ±0.5µm, which is consistent with the limitations of laser micromachining. In a production environment, optical microscopy can provide adequate measurements for in-process quality assurance, especially if used on a process validated with SEM measurements. The benefits of an inline measurement system, lower cost, and non-destructive nature, make optical microscopy an attractive image analysis technique to keep up with the continuing miniaturization of life-science devices.
 

Resonetics
www.resonetics.com

About the authors: Scott Marchand Davis is the director of quality assurance and regulatory affairs; Scott Pollock is a senior quality engineer; and Jerome DeGuzman, is a quality engineer, all at Resonetics. They can be reached at 603.886.6772 or sales@resonetics.com.