Segregation of Particulate Solids

by Majken Boesgaard Graversen,, Master of Science in Chemical and Biochemical Technology from Technical University of Denmark.


Segregation can be an issue in industry connected with e.g. handling and transportation of particulate products. The change of quality of the product due to segregation can have profound economical consequences. This communication aims to investigate the effect of particle movement by vibration (frequency, amplitude, and time), particle size distribution, and particle density on powder segregation. An experimental setup capable of extracting samples from a column of particles was constructed in order to carry out these investigations. The experiments were performed with the following materials: quartz sand, salt (Na2SO4) and non pareil sugar cores. Segregation has been observed as fine sand particles percolated through the coarse sand particles and were distributed in the whole stack. Similar results were seen with sugar-sand and salt-sand experiments. The density experiments indicated that the material density has an influence on the distribution of the sugar and salt particles.


In powder technology, different types of segregation are known depending on the powder particle size distribution (PSD), shape, density, and the forces exerted on the powder. In this article, the focus is on size and density segregation. In size segregation it is expected, that the finer particles will percolate through the coarser towards the bottom. For density segregation it is expected that lighter particles will move upwards when the bed of particles is exposed to vibration.

Experiments / Equipment:

During the design of the experimental setup several decisions were made regarding sampling, filling of column, and application of vibrations. After the powder mixture have been exposed to an external force, it should be possible to take a sample of powder from a tray in the middle of the column.

A rectangular tray with a height of 2 cm, a cylindrical hole, and sample chamber was designed, as shown in Figure 1. 12 trays were made with a lid at the bottom and top to make is possible to place the stack of trays on a sieve shaker. The design made it possible to investigate how powders move in a cylinder, and take samples from the middle of the stack removing the trays from the top and down. Sampling was performed by sliding a tray backwards making the powder fall into the sampling chamber in the tray below as shown in Figure 1.

Figure 1 – Experimental setup


Figure 2 illustrates the initial packing of the cylindrical stack in the trays used during the different experiments. Figure 2a indicates size segregation and density segregation with powder having a higher density than A. Figure2b shows density segregation where powder has a lower density than material A.

Figure 2 – Filling of segregation experimental setup.  Figure a, Filling of equipment for investigation of size and density segregation. Material A is sand, and material B is either sand or Na2SO4. Figure b shows the stack filling to investigate density segregation. Material A is quartz sand and material B is sugar particles.


For material A, quartz sand was chosen and material B was either fine sand, salt, or sugar, depending on the type of segregation studied. In the experiments with size segregation, fine sand particles were placed at the top trays expecting that the fine particles would move downwards, when the bed of particles was exposed to vibrations. For materials of same size and different density it is expected that the heavier particles will move downwards and lighter particles upwards, when the stack is exposed to vibrations. The ratio of the materials is 80 wt% of major component A and 20 wt% of minor component B of the total mass.

Results – Size Segregation For size segregation, experiments are performed with sand particles. Material A is coarse sand particles with a particle size distribution (PSD) of 710-1000µm or 500-710µm and the minor fine sand particles (material B) has a PSD of 212-250µm using the packing illustrated in Figure 2a.

Figure 3 –Comparison of quartz sand particles with PSD (500-710 µm) and (710-1000 µm) vibrated in 20 minutes at three amplitude settings; 0.2 mm (blue markers), 0.6 mm (red squared markers), and 1.0 mm (green triangle markers). The black line with filled circled markers are the starting conditions for fine particles in a stack with coarse particles of size 500-710 µm, and the black line with non-filled circle markers are starting conditions for a stack of size 710-1000 µm coarse sand particles. The x-axis is the stack height in cm. The y-axis is the weight percentage of fine particles (212-250µm).


Figure 3 illustrates a comparison of coarse quartz sand particles vibrated in 20 minutes at three amplitude settings with the fine sand particles initially placed at the top of the stack, as shown above in Figure 2a. The results indicate that the amplitude has an effect on how the fine particles sift down between the coarser particles, but not much change occurs by increasing the amplitude from 0.6 to 1.0 mm. Clearly, the lowest vibration amplitude of 0.2 mm only gives very little sifting of the small sand particles. As expected better mixing of sand particles is observed with coarse particles with a PSD of 710-1000µm compared to 500-710µm. The difference in percolation of fine sand particles may be due to the voids between the coarse particles when the PSD of the coarse particles is decreased.


Density segregation

The difference in density of two particulate solids can also cause segregation. This is studied using sugar and sand particles, and salt and sand particles.

For experiments with sugar and sand particles, the lighter sugar particles are placed at the bottom of the stack as shown in Figure 2bvarying the operating time and amplitude. At 0.2 mm for both an operating time of 10 and 20 minutes the two trays near the bottom contain only sugar, but the third contains 56.3 wt% of sugar indicating that the sugar may have packed closer together exposed to vibrations as the tray located at 6-8 cm do not contain any sugar. It can be seen that there is a difference between an amplitude of 0.6 mm vibrated for 10 or 20 minutes. The difference may indicate that settings of an amplitude of 0.6 mm vibrating of 10 minutes are not enough for the heavier sand particles to move downwards in the stack pushing the lighter sugar particles upwards. The experiment with an amplitude of 0.6 mm and an operating time of 10 minutes has been reproduced where the same tendency was seen.

Figure 4 – Sugar and sand experiments with a PSD of 250-300µm. On the x-axis is shown the stack height in cm, and on the y-axis the weight percentage of sugar particles in %.


Other studies on segregation density are performed with salt and sand particles with a PSD of 250-300µm. Salt particles have a higher density than sand and are placed at the top trays of the stack. It is expected that the salt particles will move towards the bottom.

Figure 5 – Salt and Sand experiments with a PSD of 250-300µm. The x-axis shown the stack height in cm, and the y-axis is the weight percentage of salt particles in %.


Comparing the operating time of 10 and 20 minutes shows that the time does not have a significant effect on how the salt particles are distributed in the stack. At an amplitude of 0.2 mm the salt particles do not move very far down in the stack at either 10 or 20 minutes of vibrations. This may indicate that the amplitude (force) is too low to make the salt particles move down as the sand particles are packed tighter together after the vibrations starts and some salt particles are packed in between them.  For an amplitude of 0.6 mm the distribution of salt particles alters some when increasing the operating time by 10 minutes. For an amplitude of 1.0 mm the distribution of salt particles is nearly constant at 20 wt% for both operation times. The increase in amplitude gives a more even distribution of salt particles than for 0.6 mm. Experiments with 0.6 mm the salt particles distribution vary more than 1.0 mm but has a higher weight percentage in the bottom of the stack. The difference in weight percentage for salt-sand and sugar-salt experiments could be due to the density difference between the materials.


In this study the segregation of particle solids due to size and density differences were investigated using placebo materials; sand, sugar and salt particles. For size segregation, it was observed that the fine particles sifted down through the coarser ones. It is observed that fine particles percolate more in between coarse particles with a PSD of 710-1000µm than 500-710µm as the voids between the coarse particles 710-1000µm is larger than for 500-710µm. For density segregation, using sand as the dense material and sugar/salt as light materials, the lighter particles move upwards when exposed to vibrations. Appling forces with an amplitude of 0.6 mm and 1.0 mm the particles move but at some point in the experiment the movement of lighter particles is stopped, as the distribution of either salt or sugar particles do not change much increasing the operation time. This indicates that either the forces or operation time are not enough to break the packed structure between the particles.


The author would like to thanks the following: Anker Degn Jensen (CHEC – DTU), Lars Georg Kiørboe (Department of Chemical and Biochemical Engineering)) and external supervisor Kåre Jørgensen Engsted (Novozymes A/S) for their assistance during the project.


          Majken Boesgaard Graversen, Master Thesis: Segregation of particulate Solids, 2012

          Martin Rhodes, Introduction to particle technology.: Wiley, 1998.