One of the challenges that can occur in powder handling is undesired agglomeration i.e. caking. Caking can result in various problems and in extreme cases significant downtime of the process. The types of materials that exhibit this behaviour can be found in every industry, from food, to construction to pharmacy.
Case
In the case discussed below, a client experienced caking of a flake material of a few mm in size. The material would frequently prevent discharging from storage vessels and as a result cause downtime of the process.
Apart from the unconventional shape of the particles, the material can also sublimate and therefore deposition can also happen. The goal for this project was to investigate what property causes the material to cake and how to minimize the caking behaviour.
What is investigated?
In most cases caking is formed by a chemical reaction/interaction, such as the interaction with moisture from the environment. Apart from caking the load on the material and also time is also important in order for a powder to be able to cake. To determine the root cause of caking, both caking directly is investigated and the underlying properties.
The caking investigations are performed using the uni-axial set-up. For these experiments an experimental design is needed in order to investigate the influence of various external properties on the caking behaviour of the material.
Beforehand it was unknown if there was any interaction with moisture. By performing a water vapor adsorption investigation we were able to measure if the material would take up moisture and at which relative humidity the adsorption and the desorption takes place.
uni-axial experiment.
As water vapor adsorption is measured based on the mass at different conditions, sublimations can also be measured. To confirm sublimation occurs, microscopy is used to confirm changes in surface structure over time.
Temperature can also influence the caking behaviour of materials, for example should the glass temperature of the material be close to the storage conditions it may significantly influence the caking tendency. To this end Differential Scanning Calorimetry (DSC) is performed.
Based on the above techniques an experimental design is set-up for the uni-axial experiments. These experiments are used to investigate various conditions and measure the caking strength, supported by the results of the aforementioned techniques.
Water vapor adsorption
For these types of samples static vapor sorption was preferred over dynamic vapor sorption in order to try to reduce the sublimation rate of the material, influencing the results. Both the adsorption curve and the desorption curve are measured. The adsorption curve provides information at which humidity moisture is adsorbed onto the material and desorption provides information if the moisture is strongly bonded to the material and if the material possibly sublimated during analysis.
The results display an initial decrease in mass, indicating drying or sublimation of the material. At higher RH’s (30—60 %RH) the mass stabilizes and at even higher RH the mass start increasing.
The stabilisation of the mass can indicate either that the materials has 1) fully dried, 2) starts to adsorb water and/or 3) possibly suppression of the sublimation rate due to increased water vapor pressure and water on the surface of the particle, effectively reducing the surface area that can sublimate. Both the sublimation and effect of vapor pressure was confirmed by a blank measurement.
Microscopy
Visual observation can help with understanding how the material behaves or changes over time. For this material microscopy was used to confirm that the material sublimates. The picture below show the surface of a particle.
The images shows a rather smooth surface with smaller crystals on top of the particle. At fixed intervals the microscopy camera took a picture and below is a picture of the material after 14 days.
The picture shows a much rougher surface and the small crystals in the middle are mostly gone and confirmed that indeed sublimation takes, as was suspected with the water vapor sorption technique.
Differential Scanning Calorimetry
To measure temperature dependency and also potentially detect contaminations, differential Scanning Calorimetry (DSC) is applied.
The DSC curve displays only one peak, which is the melting point of the material, indicating that should caking occur, it is not temperature dependent and no contamination is present.
Uni-axial
Based on the results of previous techniques various condition are investigated. For the various conditions a baseline conditions was selected: 20°C /
55 %RH. In the table below the different conditions investigated in no specific combination are shown.
Due to the unconventional shape of the particles the results were not very repeatable as creating the same packing was not possible, resulting in sometimes relative standard deviations of more than 50%. The experiments were therefore at least performed as independent duplicate.
The big-bags in which the material is stored was stacked two big-bags high. To investigate of the load of two big-bags influence the caking behaviour the load was variated. The results show that the load influences the caking behaviour, especially during longer storage periods the unconfined failure strength increases.
The different variations of the parameters
investigated of the material.
| Time (h) | Temp. (°C) | RH (%) | Load (big-bags) |
|---|---|---|---|
| 1 | 5 | 15 | 1 |
| 24 | 20 | 55 | 2 |
| 72 | 40 | 88 | |
| 168 |
Temperature was also investigated and revealed that the unconfined failure strength would increase the with temperature. Based on literature the partial pressure of the material also increases with temperature and in a similar order as the unconfined failure strength, again an indication that sublimation places a role in the caking behaviour.
Conclusion
It was confirmed that the samples indeed cake and a combination of factors influences this behaviour. The load, storage time and temperature on the material can significantly increase the caking strength.
The effect of relative humidity on the caking strength was in the end for this material not significant.
Even though we were able to confirm which parameters influences caking it was not enough to prevent caking. It was suspected that the flake particles created a powder bed like a brick layers that could easily support a large load The material was therefore produced as trial via prilling, creating spherical particles instead of flake particles. The spherical particles have much smaller contact areas and resulted in minimal to no caking. Showing that in the end the shape was the most important factor.
