Comparison of Particle Size Measurement Methods: Dynamic Image Analysis vs. Laser Diffraction vs. Sieve Analysis

Different Measurement Techniques, Different Results—So, Which One Is Correct?

The most commonly used methods for determining particle size are Dynamic Image Analysis (DIA), Laser Light Scattering (Laser Diffraction), and Sieve Analysis. In this article, you will find these techniques and comparisons between them. Each method has different measurement ranges, and the characteristic sizes at which measurement is possible are shown in Figure 1.

This article will help you decide which method is more suitable for which applications in particle analysis. The devices used in this article are the Retsch Technology Camsizer P4 for Image Analysis, the Horiba LA-960 for Laser Diffraction, and the Retsch AS 200 for Sieve Analysis. Sieve Analysis: The Traditional Method

Sieve analysis is the traditional and most widely used method for determining particle size. A sieve set consists of several sieves with increasing aperture sizes, and the sample is placed on the topmost sieve. The sieve set is placed in a shaking device and vibrated for a specified duration. As a result, particles are distributed according to their sizes and the mesh openings of the sieves (fractions). Ideally, particles pass through the smallest sieve opening with the smallest projection area possible. Taking cubic particles as a model, this corresponds to the edge length of the cube. For irregularly shaped particles, the size determined by sieve analysis will be a value between thickness and diameter, since the particle is oriented diagonally through the sieve hole (Figures 2 and 3). For this reason, sieve analysis is a technique that preferentially measures particles in the direction that tends to determine particle width. Sieve analysis is performed until the residual mass of the sample on the relevant sieves no longer changes (constant mass). Each sieve is weighed individually, the volume of each fraction is calculated as a percentage by mass, and a mass-related distribution is obtained. The result of sieve analysis is limited by the size of the fractions used. A standard sieve set consists of a maximum of 8 sieves, which means the particle size distribution is based on only 8 data points. The process cannot be automated, making it time-consuming.

The steps of sieve analysis are as follows:

Initial weighing, 5–10 minutes of sieving, reweighing, and sieve cleaning. Common errors encountered in sieve analysis are: sieve overloading (blocking of mesh openings, overly coarse results); old, worn, or damaged sieves (overly fine results); or data transfer errors. It should also be noted that the aperture sizes of new sieves compliant with standards are subject to certain tolerances. For example, the average actual aperture size of 1 mm sieves is permitted a deviation of approximately ±30 µm, and for a 100 µm sieve ±5 µm (meaning the average actual aperture size is between 95 and 105 µm). However, this is merely an average value, meaning only some openings may be even larger. With sufficient sieving time, particles find the largest openings through the sieve, resulting in particles larger than the nominal aperture size indicated by the sieve. Thus, the sieve appears larger than its nominal size indicates. These tolerances are particularly evident in sieve analysis results for spherical samples or samples with narrow particle size distributions, as clearly shown in Figure 4, which displays the measurement results of a glass bead sample. Measurements can only be compared through points representing the sieve fractions. As can be seen, there is excellent agreement between the results. Looking more closely at the data at 710 μm, the deviation of the Q3(x) value is seen to be 6%. While this may seem quite large at first, the deviation in size is only 13 μm, and this deviation falls within the 710 μm sieve tolerance. Because the cumulative curve is very steep at this point, a small difference in size has a strong effect on the Q3(x) value.

Dynamic Image Analysis (DIA): You Measure What You See

There are two image analysis techniques for particle characterization. Static image analysis is essentially a microscope that measures a sample placed on an object slide step by step. Although the image quality is very good and optical resolution is quite high, this method has some decisive disadvantages in terms of presenting particle size distributions. The size range is limited, the process is quite time-consuming, and the quantity of particles analyzed is usually insufficient to obtain a statistically robust statement about the entire sample. As a result, in this article we will only discuss dynamic image analysis. This technique involves a camera system in front of an illuminated background and a particle stream passing in front of the camera. Figure 5 shows a schematic of this measurement principle as applied in the CAMSIZER® X2. The system measures freely falling particles and suspensions and also demonstrates the dispersion of particles prone to agglomeration by means of air pressure. Modern DIA systems analyze more than 300 images per second in real time, detecting millions of individual particles in just a few minutes. This performance is based on fast cameras, bright light sources, short exposure times, and powerful software. Unlike sieve analysis, DIA measures particles in a completely random manner. Based on particle images, size and shape parameters are determined. Typical size parameters include, for example, the width, length, and diameter of an equivalent circle (see Figure 6). Parameters describing particle shape include sphericity, symmetry, convexity, and aspect ratio. An important characteristic of DIA is its extremely high detection sensitivity for large-sized particles. For example, CAMSIZER® P4 is designed to detect every individual particle in a sample. For large-sized particles, the CAMSIZER® X2 model has 0.1% accuracy. The resolution of DIA systems is also unmatched. The smallest size differences in the micrometer range are reliably detected, and multimodal distributions are determined with absolute accuracy. As a result, different distributions can be produced for the same measurement. In this example, the red curve is based on measuring particle width. Blue represents a particle length. The X-area parameter specifies the diameter of the equivalent circle defined as the particle size. The results ultimately depend on the original question to which they are relevant. When examining fibers or extrudates, length parameters are of interest. If comparison with sieve analysis is required, width is more important. For DIA compared to sieve analysis, "width" is the common parameter. However, when measuring irregularly shaped particles, there are systematic differences in the results obtained because DIA measures particles in random orientations. Figures 2 and 3 illustrate how differences in particle size measurement arise and how they can be interpreted, using special particle shapes. The differences in particle size distributions are systematic for each defined particle shape. CAMSIZER® software contains algorithms that correlate DIA results with those obtained from sieve analysis to approximately 100% (Figure 7).

This procedure is frequently applied in particle size analysis applications for quality control, because in a globalized market many products are analyzed by different laboratories using different measurement techniques, and thus the need for comparability arises. Laser Diffraction: Spheres and Correlations

In the static laser light scattering analysis method, also called laser diffraction, particle size is measured indirectly by detecting the intensity distributions of laser light scattered by the particles. Figure 8 shows the setup of a modern laser particle size analyzer such as the Horiba LA-960. This method is simply based on the principle that large particles scatter light at small angles, while small particles produce wide-angle scattering patterns. Large particles produce a very sharp intensity distribution with distinct maxima and minima at defined angles, while the light scattering pattern of small particles gradually disperses and the overall intensity decreases. Because the individual light scattering signals of particles overlap, measuring multimodal particles containing particles with different distributions is particularly difficult. Figure 8. Laser light scattering spectrometer Horiba LA-960 uses two light sources and 93 measurement channels to record scattered light images at wide angles. It is possible to measure emulsions and dry powders in the range of 0.01 nm to 5,000 μm. Static laser light scattering (SLS) is an indirect method that calculates particle size distributions based on the scattered light patterns produced by particles. The algorithms are based on the well-known Mie theory, assuming that particles are spherical and that optical properties such as refractive index (RI) and absorption index (AI) are well known. The greatest advantage of SLS is its wide measurement range. None of the other techniques discussed in this article can reliably detect particles smaller than 1 micron. The SLS method is quite user-friendly and can be largely automated. The only disadvantage of this method is that it has relatively poor resolution. Even the latest analyzer cannot detect large size fractions very precisely if their volume is below 2 vol%. More than three different components in a mixture essentially cannot be detected. Figure 9 shows an example of a mixture of polystyrene-latex standard particles. Dynamic image analysis can distinguish all four different particle sizes exactly, whereas the laser diffraction analyzer cannot correctly resolve 10 μm and 12 μm particles. Figure 9. Measurement of a mixture—Four particle standards (2.5 μm–5 μm–10 μm–12 μm). While DIA can distinguish four components (red), laser diffraction distinguishes only three components. Figure 10 demonstrates the comparability between SLS, DIA, and sieve analysis using a ground coffee sample. Sieve analysis gives good results. Measurement of particle width with CAMSIZER® X2 gives results closest to sieve analysis. There is no clear comparability between sieve analysis and laser diffraction; the result obtained with SLS roughly corresponds to the X-area parameter (diameter of the equivalent circle). All measured various particle sizes are correlated with spherically shaped particles. For this reason, SLS always produces broader size distributions than image analysis. Figure 10. Measuring ground coffee using different methods. DIA, particle width (red); DIA, particle length (blue); DIA, equivalent circle diameter (green); Laser diffraction (orange*); Sieve analysis (black*). The measurement comparison of cellulose fibers in Figure 11 makes the comparison of methods even clearer. DIA sharply distinguishes between fiber thickness and length, whereas SLS cannot. The laser diffraction measurement curve is first parallel to DIA's width measurement (red) and then approaches "fiber length" (blue). As a result of laser scattering, width and length information are obtained and all are combined into one size distribution. Figure 11. Measurement of cellulose fiber. CAMSIZER® XT (image analysis) measures particle width (red), particle length (blue), and X-area (green). SLS measurement (*) is a mixture of width and length and shows a continuous transition. DIA, width and length.

Conclusion

Of the methods presented in this article, dynamic image analysis provides the most precise information about particle size because it also takes shapes into account. This means that results obtained with other methods can be tracked and, if necessary, made comparable. Through direct measurement of particles, the information content and resolution of DIA is far more pronounced than laser diffraction and sieve analysis. The advantages of sieve analysis are its traditional widespread use and relatively low equipment cost. Particle size analysis by laser diffraction is a reliable technique capable of measuring sizes below 1 micron. Caner Pekel / Chemical Engineer / Kutay Laboratuvar Cihazları