Application Note: PrismLink™

The surface texture of microscopic particles and powders can serve as a valuable indicator of properties such as reactivity and origin and the physical performance of these particles and powders in applications like coatings and additive manufacturing. Such particles are frequently subjected to analysis in the Scanning Electron Microscope (SEM), which is an excellent tool for quantifying particle morphology (e.g., area, diameter, perimeter length, perimeter fractal), and shape descriptors such as roundness, form factor, and elongation. The SEM’s ability to use automated data acquisition, such as IntelliSEM, allows for rapidly collecting thousands of high-resolution 2-dimensional images with sharp edge definition, facilitating the rapid characterization of large populations of particles using these types of morphological measurements.

Unfortunately, these conventional morphological descriptors, as defined by ASTM Method F1877, typically do not adequately characterize surface texture. The measurement of surface structure is a difficult proposition in a fundamentally 2D instrument. While depth measurements are possible with the SEM using 3D reconstruction techniques such as stereo-pair merging, tomography via physical slices, and Focused Ion Beam (FIB) sectioning, these methods do not lend themselves to the rapid analysis that is required for generating statistical profiles of large populations.

PrismLink™ uses the 2-dimensional backscatter (BSE), secondary (SE), or mixed electron signals to rapidly compute a single descriptor that estimates the relative surface roughness of a particle during an IntelliSEM analysis.

Figure 1 shows two particles, which the human eye readily identifies as “smooth” and “rough”.

Figure 1: Visually “smooth” and “rough” particles - BSE mode SEM images (top) and intensity-modulated renderings (below)

However, the perimeter-based morphological measurements performed on SEM images will give similar values for both particles, due to their similar edge characteristics. Table 1 illustrates selected perimeter-based measurements, including Aspect Ratio, Form Factor, and Roundness, in addition to PrismLink measurements for the two particles in Figure 1.

As expected, the Form Factor, being proportional to area / perimeter², shows the most pronounced variation with particle roughness. However, none of the perimeter-based properties are as effective in capturing the differences in surface structure as PrismLink; note the significant difference in PrismLink values:

Table 1: Comparison of perimeter-based and PrismLink measurements for the particles in Figure 1

Using intensity as a surrogate for the third dimension, PrismLink™ measures the surface texture of the particles and generates a single measurement for each particle. Figure 2 shows PrismLink measurements (in red) for various similar-sized particles. In this figure, the top-left particle serves as an “exemplar of smoothness,” and the relative PrismLink – the ratio of the PrismLink value for each particle to that of the exemplar – is shown in green.

Figure 2: Examples of PrismLink™ measurements on a variety of particles, along with Relative values when compared to top left particle, which serves as “exemplar of smoothness”

In addition to the conventional morphological measurements and elemental compositions, PrismLink adds a texture-specific metric that can aid in the identification and classification of particle species, thereby providing a more comprehensive understanding of the properties and performance of the materials under study.

Case Study

A manufacturer of materials for use in harsh and extreme environments recently observed unexpected performance variations in additively manufactured components. Despite all powders used in the production meeting the required particle size distribution specifications, slight differences in their particle size distributions were noted. These variations, which had not previously affected the manufacturing process or the quality of the components, are now raising concerns about their potential impact on performance.

Figure 1 - Particle size distributions of 4 powders; components printed with powder lots 10605253 and 10605254 failed to meet end user specifications

The use of image-based particle size distributions can facilitate a more profound understanding of the distinctive attributes of individual particles. Modern image-based particle characterization techniques, including Dynamic Image Analysis, light optical microscopy, and IntelliSEM™-based scanning electron microscopy, can rapidly provide statistically significant details on particle features.

The form factor, which is sensitive to variations in the perimeter, can reveal differences in the distribution between two powders. As illustrated in the figure 2, the form factor indicates differences between the powder lots; however, it does not fully explain the underlying causes of these differences. For instance, these discrepancies may be attributed to variations in the number of satellite particles or surface texture as shown in Figure 3.

Figure 2 - Form Factor distribution for the 4 powders. A spherical particle has a form factor of 1.

Figure 3 – Images of all particles from sample 10605253 with Form Factor of 0.7. Particle identifier is shown in white in each image.

Figure 3 shows all particles from sample 10605253 with a Form Factor of 0.7. Some particles are misshapen such as particle #39 whereas others have satellites such as particle #1275 or have rough surface structure such as particle #31. Clearly, a distribution plot of the Form Factor shows differences between these two batches of powders yet it doesn’t provide insight into what drives this difference.

A plot of the PrismLink highlights both similarities and differences between the two powders. Approximately 37% of the particles in each powder exhibit the same surface texture. However, the remaining 63% show significant variation: in regular powders, this portion demonstrates a gradual increase in surface texture, whereas in the newer powders, these particles have a distinctly rough surface.

Figure 4 - Distribution of PrismLink values for the 4 powders. A higher PrismLink value indicates higher surface texture. The dotted line represents 37% of the particles in the powder.

Summary

Surface texture is a critical parameter in numerous chemical reactions, affecting reaction rates and outcomes. Conventional techniques, such as Dynamic Image Analysis, typically infer surface texture based on the particle's perimeter. We present a novel parameter, PrismLink, which directly captures the surface texture of particles. This parameter immediately highlights discrepancies between batches of particles that were presumed to be identical, offering valuable insights for additive manufacturing.