Digital Particle Analysis

New digital imaging techniques are providing crucial information about the shape and performance of abrasive particles used in grinding and polishing operations.

Achieving the optimum results from a grinding or polishing process starts and ends with choosing the right abrasive. And while the material properties of the right abrasive are important (hardness, for example), the particulate properties are also critical. After all, abrasives are a collection of a large number of particles. It is the statistical average of these particulate properties that defines most of the performance characteristics of the abrasive.

Methods such as laser diffraction and sedimentation are commonly used to measure some of these particulate properties.¹ However, these techniques are only able to provide general information-specifically, mean diameter and range-about the size distribution. While such information is useful, it does not present a complete picture of the mechanical properties of the abrasive particles. This is because the particle sizing methods used to characterize abrasives measure particle size indirectly.

For example, laser diffraction, which is often used to obtain the particle size distribution of abrasives,² actually measures the spatial (angular) variation of light that has scattered off an ensemble of particles. The scattering pattern is then curve-fit or otherwise mathematically transformed to produce a particle size distribution. The mean diameter and range provided are actually referenced to an equivalent sphere. This means that the best approximation to the true particle size these methods can provide is the size of a sphere with similar scattering properties. While some grinding abrasives are made up of spherical particles, most abrasive particles have very irregular shapes, as shown in the microscope image in Figure 1.

As it turns out, these irregular shape characteristics have been known for a long time to be extremely important with regard to the performance of an abrasive material.³ Furthermore, this information cannot be obtained with the common particle sizing methods mentioned earlier, but can only be determined by image analysis. Until recently, this meant microscopy. However, with its attendant multi-step sample preparation and optical adjustments, the measurement and recording of shape information using microscopy is such a time-consuming task that most engineers would rather rely on the incomplete characterization provided by sieves or particle size analyzers.

Fortunately, recent advances in digital imaging and the large amount of computational power available on personal computers now makes it possible to bring shape analysis into the quality control laboratory and onto the production line. This has led to the development of new ways to quantify and present shape information, and new ways to characterize and understand grinding and polishing processes.

Particle Analysis Parameters

To appreciate the discriminating capabilities of shape analysis and how it applies to abrasive materials, consider the collection of particles shown in Figure 2. The particles are actual digital images captured using an image analyzer. It is obvious that each of these particles has different shape characteristics and will presumably act differently if grouped together with many similarly shaped particles and used in a grinding/polishing process.

Table 1 contains various size and shape parameters calculated from these digital images. The sieve diameter is based on the diameter of the maximum sphere that can be inscribed into the particle. This has been shown to be a more robust method of assigning a diameter to a complex-shaped particle compared to the equivalent diameter,⁴ and excellent agreement with actual sieve results has been obtained.

The equivalent diameter is the diameter of a sphere that has the same area as the particle and is the more common way of assigning diameter to a particle. However, this approach does not typically provide good agreement with sieve results.⁴

The elongation is calculated by a new method in which the maximum and minimum inertial ellipse axes are determined. This is in contrast to the more common method of using the ratio of the maximum and minimum Feret diameters to determine elongation.

The wear factor, an innovative new parameter, is a measure of local curvature and is calculated from the set of maximum inscribed spheres. The larger the percentage, the more blunt asperities are present in the particle shape. This approach is superior to the commonly calculated "shape factor," which is a ratio of the particle's area over the particle's perimeter. The "shape factor" does not differentiate well between particles that are blunt vs. particles with sharp edges.

As can be seen in Table 1, all of the particles have the same sieve diameter, which means that they would all be held in a 120-mesh sieve. Thus, it is not likely that sieve analysis will allow the different abrasive properties of these particles to be determined.

The elongation covers a range from compact (particle 4) to high aspect ratio (particles 1 and 3). However, the most important parameter in differentiating these particles is the wear factor, which tells us that particles 1 and 2 have more sharp edges than particles 3 and 4. Grinding/polishing materials composed of particles that have many sharp edges will tend to remove more material or otherwise grind faster than a material made up of particles with smooth surfaces. Thus, it should be expected that abrasives made up of particles with similar wear factors, like those of particles 1 and 2, would perform differently. Abrasive particles with too many sharp edges might gouge a surface during polishing or grind at faster rates than expected.

It is important to note that none of the other size and shape parameters correlate with the wear factor. This example illustrates how particle shape analysis can provide critical information about abrasive performance-information that cannot be measured by sieves or particle size analyzers.

New Image Analysis Techniques

There is more to the measuring of particle shape than the new morphological parameters described in the previous section. Digital imaging has emerged as the enabling technology that has added new particle characterization methods to the arsenal available to process engineers in charge of a grinding/polishing process. These new methods not only enable the quality control laboratory to measure particle shape, but in some cases can also allow companies to monitor this information in-process. As a result, it is possible to monitor the wear factor, for example, in situ and determine when the abrasive particles start to wear down and become less effective.

The ability for in-process measurement is just one way that these new imaging methods stand out from classical microscopy. In-process measurement is possible because of the ability of these new instruments to capture and digitize the images of many particles in a short period of time. Having the images of many particles enhances the statistical accuracy of the analysis. Also critical is the ability to quickly process all the digital images and calculate various size and shape parameters.

The key to all of these enhancements is the automation allowed by computer control, both in terms of capturing the images and turning them into useful information. This new automated image analysis is performed through either a dynamic or static method.

Dynamic image analysis involves flowing the particles, either in an air or liquid stream, past a camera used to capture the images. The advantages of this method are that many particles can be brought into the focal plane of the camera with little effort, and that the image analysis can be performed in-process.

The downside to this approach is that it is difficult to control the particles, both in terms of their orientation and the number of particles that end up in the frame (particularly in-process). Since the flow path is often wider than the depth of field, it is possible to lose details or resolution because the particles are out of focus. This usually limits the technique to particles no smaller than about 10 microns. Also, it is difficult to disperse fine "sticky" powders well enough (especially in an air stream) to be sure that the system is imaging individual particles and not agglomerates.

Static image analysis entails the imaging of particles placed on a slide. In the case of freely flowing powders, this operation can be automated by feeding particles onto a moving optical surface, thus enabling the possibility of on-line measurements. In the case of a fine or "sticky" powder, compact and easy-to-use dispersion devices can be used to place the particles on the slide in a way that breaks up agglomerates and keeps them from overlapping each other.

Coupled with a computer-controlled X-Y translator that moves the slide past the camera, static image analysis allows hundreds or thousands of particles to be imaged in minutes. Dispersing particles on a slide ensures that they will be oriented in one way and always in focus. This leads to higher resolution and thus more detail than can be attained by dynamic image analysis. Of course, the application of particles to a slide means that a much smaller amount of material can be tested than with dynamic image analysis, which can limit the statistical accuracy of the results.

Better Control

Microscopy has always been an important tool to study particles, but up to now it has required highly trained operators and a lot of time to perform. Image analysis does not make traditional particle size analysis obsolete-sieves can still test huge aliquots of material, and particle size analyzers can provide size information that averages together millions of particles compared to the hundreds or thousands that these new image analyzers can capture. But the new digital dynamic and static image analyzers can now be used in conjunction with the particle size analyzer or sieve stack, providing complementary information that will allow process engineers to better control their grinding and polishing processes.

For more information about particle characterization, contact Particle Sizing Systems, 8203 Kristel Circle, Port Richey, FL 34668; (727) 846-0866; fax (727) 846-0865; e-mail pohagan@pssnicomp.com; or visit http://www.pssnicomp.com.

References

1. O'Hagan, P., et. al., "Shedding New Light on Grinding and Polishing," Ceramic Industry, Vol. 154, No. 11, 2004, pp. 14-16.

2. Thornton, A., "Ensuring Quality in Particle Size Distribution Analysis," Ceramic Industry, Vol. 153, No. 4, 2003, pp. 45-50.

3. Krumbein, W., "The Effects of Abrasion on the Size, Shape, and Roundness of Rock Fragments," Journal of Geology, Vol. 49, 1941, pp. 482-519.

Pirard, E., et. al., "Direct Estimation of Sieve Size Distributions from 2-D Image Analysis of Sand Particles," in Proceedings of Partec 2004, Nuremburg, Germany.