As appeared in April 2014 PBE Copyright CSC Publishing www.powderbulk.com I n this column, we’ll examine segregation, a major cause of blending failures, as we continue the discussion of batch blending failure modes in the last two “Mixing Mechanics” columns (April and September 2013). Segregation (also known as demixing) commonly occurs during processing of powder blends that contain ingredients with even slightly different particle sizes. While segregation can occur in various situations, in every case it leads to the same — and usually undesirable — outcome: As the powder blend flows through the process, the blend becomes less homogeneous and displays increasingly varied composition. In the last 20 years, researchers who study bulk solids behavior have proposed many mechanisms to explain why blend segregation occurs in different situations. However, all known blend segregation mechanisms share a common cause: In a powder blend, free-flowing particles or particle agglomerates that have different properties — especially size and density — will follow different trajectories as the blend flows. The different trajectories separate the particles as they move downstream, causing the composition, particle size and particle size distribution, and other properties of different portions of the blend (and thus, different units of the end product) to fluctuate substantially. This phenomenon, known as differential motion, helps explain why and how segregation takes place. Let’s explore the most common scenarios for blend segregation: inside the blender, during blender discharging and filling of downstream receiving equipment, and during sampling. Segregation inside the blender While a blender’s purpose is to increase blend homogeneity, segregation can and does happen inside the blender. Different batch blender types exhibit different forms of segregation. Fernando J. Muzzio Practical powder blending: Blend segregation Most tumbling blenders, such as the V-blender (also known as a twin-shell blender or Patterson-Kelley blender), double-cone blender, and bin blender, provide repetitive rotation around one axis. As described in the September 2013 column, the tumbling blender provides fast convective mixing in the powder’s motion direction (typically, around the blender’s rotation axis) but very slow dispersive mixing across the powder’s motion direction (typically, axially — along the blender’s rotation axis). As the powder tumbles around during blending, particles of different sizes tend to segregate along the blender’s rotation axis. This phenomenon produces axial bands of different particle sizes in the rotating blender. In contrast, a convective blender, such as a ribbon blender or plowshare blender, typically has a rotating impeller or blade along the blender’s axis and imparts a more complex flow, with convection as the dominant blending mechanism in all directions. As a result, this blender tends to produce less intense segregation, except in a dead spot (a zone where the impeller fails to agitate the blend), where larger and denser particles tend to accumulate. In both tumbling and convective blenders, another segregation mechanism — fluidization segregation — can occur. As the blender operates, the powder becomes aerated and reaches its minimum bulk density. When the blender is stopped, the blend settles, releasing some of the excess air and dragging the blend’s finer particles along with it. This yields a vertically stratified blend, in which the upper layers have a higher concentration of finer particles. Segregation during blender discharging and filling of downstream receiving equipment Segregation occurs even more commonly when the blend is discharged from the blender into a downstream drum, bin, hopper, or chute. As the powder flows from the blender into a receiving container, it typically forms a heap. Particles rolling down the heap’s slope often exhibit sifting segregation, in which smaller particles fall vertically into the interstices between larger particles and concentrate near the heap’s center. Larger particles can’t fall straight down to the same extent and instead roll downhill, accumulating around the heap’s perimeter. When the powder heap in the container is then emptied into the next process unit, the blend’s composition and particle size distribution fluctuate. We regularly observe this phenomenon in our daily lives, as in the container of bleu cheese crumbles shown in Figure 1, where the small crumbles have settled down between the large crumbles. Fluidization segregation can also occur, just as in the blending operation itself, when the discharged blend is loaded into a chute. As the powder fills the chute, it must displace the air in the chute. This creates a countercurrent flow of powder and air. The air can drag smaller, lighter particles vertically upward, creating significant fluctuations in the blend’s composition. Copyright CSC Publishing In both cases, if the powder in the receiving equipment (whether a container or chute) is further processed without reblending it, the end product’s composition often fluctuates. Consider a blending process for pharmaceutical tablets where the powder blend segregates in a hopper downstream from the blender: Every time the hopper is filled and emptied, a “concentration wave” is created in the blend flowing to the tableting machine, causing significant composition fluctuations that will affect the finished tablets’ product quality or cause tableting problems. Segregation during sampling Another common form of segregation results from using a sample thief (or thief sampler) to extract a blend sample from the blender. This device has an inner cylinder with sample cavities, each with a hole in the cylinder wall, and an outer cylinder, also with holes and a pointed tip. Before the thief is inserted into the powder, the inner cylinder is twisted so its holes don’t align with the outer cylinder’s holes; when the thief reaches the proper location in the powder bed, the inner cylinder is twisted again so that the inner and outer cylinder holes align, which allows the powder to enter the sample cavities. The inner cylinder is twisted again to close the holes before the thief is removed. One problem with using this device is that as the thief is inserted into the powder blend, particles are dragged along with it, which segregates the powder Figure 1 Sifting segregation in container of bleu cheese crumbles along the thief’s path. A bigger problem is that opening the thief’s cavities to capture samples can cause differential motion in the blend, so the resulting samples don’t truly represent the blend’s composition at the thief location. Thief sampling tends to undersample larger particles (greater than 600 microns) because the larger the particles, the lower their concentration in each sample. In an extreme case, samples with 10 to 30 percent fewer large particles than the blend actually contains can be consistently extracted from an otherwise homogeneous blend. Minimizing blend segregation Take the following steps to help minimize segregation problems in your batch blends. Correctly diagnose the cause. Start by correctly diagnosing the problem. Blend segregation is driven primarily by particle size and density differences among ingredients in the blend, and most segregation problems occur in blends with substantial amounts of free-flowing ingredients. To determine how likely your blend is to segregate, run tests of blend samples with a segregation tester. (Several models are commercially available; they can also be accessed in independent or university labs.) The instrument quantifies the intensity of sifting and fluidization segregation mechanisms in each sample. If the results show that your blend tends to segregate, you may be able to reduce or minimize this tendency while maintaining acceptable flow properties by adjusting the particle sizes of ingredients or the blend’s cohesiveness. Consider your process. Next, examine your process and determine whether you can modify it to minimize the blend’s segregation tendencies. This can involve changes such as eliminating hoppers or modifying their cone angles to promote mass flow, reducing the number of blend discharges the process requires, or adding another blending step just prior to the blend’s final processing. However, while such process modifications are often effective in reducing segregation’s impact on product quality, they don’t remove the problem’s root cause: the intrinsic tendency of free-flowing blends containing ingredients with different particle sizes to segregate. Convert from batch to continuous processing. If modifying your batch process can’t minimize blend segregation, you can take a more radical step: converting your batch process to a continuous one. A properly designed continuous process that achieves nearplug-flow material behavior after blending will typically display minimal, if any, segregation. While this less common approach to minimizing segregation has additional benefits, such as better process controllability and a smaller equipment footprint, it also requires a major change to your existing process and is only effective if you manufacture a relatively large quantity of product each year. Next time I’ll focus on a very different source of blend homogeneity problems, also driven by ingredient properties: ingredient agglomeration. PBE Fernando J. Muzzio is director of the National Science Foundation’s Engineering Research Center on Structured Organic Particulate Systems (http://er cforsops.org/) and Professor II, chemical and biochemical engineering, Rutgers University, Piscataway, N.J. He can be reached at 732-445-3357 (fj email@example.com). He earned his BS in chemical engineering at the University of Mar del Plata, Buenos Aires, and his PhD in chemical engineering at the University of Massachusetts, Amherst. He has published more than 200 peer-reviewed papers on mixing and blending, presented at numerous conferences, and earned several patents. The author will answer your questions in future issues. Direct questions to him at firstname.lastname@example.org or via the Editor, Powder and Bulk Engineering, 1155 Northland Drive, St. Paul, MN 55120 (toneill@cscpub .com).
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project