Elsevier

Powder Technology

Volume 300, October 2016, Pages 146-156
Powder Technology

Understanding and preventing agglomeration in a filter drying process

https://doi.org/10.1016/j.powtec.2016.03.003Get rights and content

Highlights

  • Undesired agglomeration can occur during agitated filter drying processes

  • Effects of process conditions on the formation of agglomerates and their respective mechanisms is different for each powder

  • Condensate drips may also lead to the nucleation and growth of agglomerates

  • Surface tension and particle solubility in wash solvent affects the formation of agglomerates

Abstract

The occurrence of severe agglomeration during a filter drying process is a challenging issue in the pharmaceutical industry and has yet to be fully understood. Product degradation, extended drying times, additional equipment required for the elimination of lumps and downstream processing issues are some of the problems caused by this phenomenon and there is great interest among researchers from various industries to explore how the filtration/drying processing step impacts the agglomeration of powder. This paper investigates the effect of operating condition such as the drying temperature, agitation speed, fill volume and blow-down period on the formation of large agglomerates. In addition, other potential sources of agglomeration such as the dripping of condensates as well as surface tension of wash solvents were also explored. A series of systematic experimental work was carried out using sodium bicarbonate, calcium carbonate and an API intermediate which has high agglomeration tendency to understand behaviour of agglomeration among the different types of powder. It was found that similar trends may be observed in the formation of lumps for different types of powders but the underlying mechanisms of agglomeration were different. The unique behaviour of each powder suggested that a universal solution or mitigation method to eliminate the formation of agglomerate may not be possible. However, by identifying the specific agglomeration mechanisms present in a powder-solvent system, it is possible to carry out the mitigation in a more effective way to minimize the agglomeration.

Graphical abstract

Systematic investigation of an agitated filter drying process for the understanding and elimination of undesired agglomeration

Image 1
  1. Download : Download full-size image

Introduction

The production of active pharmaceutical ingredients (API) often involves a series of complicated process steps which may include chemical transformation step, purification, filtration, drying and milling. Drying is often conducted as one of the last stages of the API production following the filtration step due to the need to remove the solvent from the solid phase in order to meet the residual solvent specification [1], [2], [3]. In fact, drying is one of the most crucial steps where a poorly executed moisture removal step can have serious impact on the product performance, safety as well as the compliance with quality requirements [4]. Depending on the product specifications and safety requirements, the drying of API can be carried out in various types of equipment such as fluidized bed dryer, tumble dryer, filter dryer, conical dryer etc. However, in many pharmaceutical companies, the agitated filter dryer (AFD) is one of the most popular dryers in API production especially when dealing with highly toxic substances [4], [5], [6].

An AFD integrates both filtration and drying process into a single unit operation, unlike many other conventional drying methods which requires the transfer of product from filtration equipment to a dryer. This provides advantages over other drying methods such as, minimizing personnel exposure to hazardous APIs and improved product quality due to its well contained environment [6]. From a technical point of view, an AFD offers high content uniformity due to agitation and is highly versatile, owing to its ability to incorporate various drying modes (such as cold or hot air blowing, pressure or vacuum drying) depending on the specification requirements [4], [6]. Nevertheless, the preferred choice and advantages of AFD in drug production is often complemented by several drawbacks. One of the most notable disadvantages is the susceptibility of the powder to undergo agglomeration which has been a long term issue in the pharmaceutical industry. Formation of hard, unbreakable lumps and product degradation due to the moisture content present in the lumps are two of the common downstream issues caused by this phenomenon.

Fig. 1 provides a schematic of a typical AFD and its operation flow chart. The process of agitated filter drying generally starts by charging the slurry or suspension (which is usually a product of a crystallization process) into the AFD. The slurry is allowed to settle in order to form the powder bed prior to the removal of solvent via pressure filtration. The filtered cake is then washed with wash solvents such as water, or 10% solutions of ethanol, acetone and 2-methyltetrahydrofuran to remove the residual mother liquor or solvent left behind from the upstream process. The choice of wash solvent will depend on several factors such as the solubility of main product and the target impurities as well as the effectiveness of the solvent in displacing the residual mother liquor. In general, there are two modes of filter cake washing step namely, (1) displacement wash, and (2) re-slurry wash. A displacement wash is carried out by spraying the wash liquid from the top of the vessel via a spray bar which creates a plug flow motion that pushes the mother liquor out through the bottom of the vessel. One the other hand, a re-slurry wash is carried out by filling up the vessel with large amount of liquid while stirring the re-suspended solids in order to dilute the impurities or mother liquor present in the filter cake. At the end of the washing sequence, pressure filtration is carried out again and is extended over a longer period of time (usually for a few hours). This is also known as the blow-down or blow-through step which is aimed to remove as much residual moisture as possible prior to the agitated drying step [7], [8]. Once the blow-down step is completed, drying of powder is carried out at the desired jacket temperature and vacuum level while the agitator is lowered to mix the powder bed for better heat transfer.

In recent years, various studies had been undertaken to investigate how the operation of an agitated drying process affects the agglomeration behaviour of powder during the process. For instance, Lekhal et al. studied the impact of agitation speed, pressure and drying temperature on the drying of potassium chloride and L-threonine crystals in a lab scale agitated dryer [1], [2]. They showed that there was a significant decrease in the final average diameter of the crystals associated with an increase in agitation speed. This was attributed to the greater interparticle and particle-impeller collisions which ultimately contribute to the breakage of crystals. They also reported that drying at higher temperature will lead to an increase in particle size of the crystals due to a faster drying period and reduced shearing time. Furthermore, the increase in particle solubility as a result of increased drying temperature will lead to strong connections between crystals that will contribute to the strength of the agglomerates [1], [2], [9]. In addition, Hsu and Chiou had demonstrated that drying alumina powders under vacuum at higher drying temperature of 100 °C tends to create uneven drying of the powder which will lead to more severe agglomeration [10]. Similarly, Sahni et al. investigated the effect of process parameters such as wall temperature, impeller speed and fill level on the drying performance of lactose monohydrate in a lab scale filter dryer [3]. Lactose experienced a slight increase in the particle size distribution with increasing wall temperatures. This was associated with reduced attrition or shearing time, in line with the work by Lekhal et al. [1], [2]. A significant decline in the particle size of the lactose monohydrate was also observed as the agitation speed was increased. On the other hand, the authors also showed that there was a significant increase in the particle size when the fill volume of the vessel was increased. The phenomenon has been ascribed to the reduction in the interparticle and particle-impeller collision as not all particles are exposed to the shearing action of the impeller blade. Other AFD related studies which investigates the impact of drying conditions on the particles properties have also been conducted [11], [12], [13], [14], [15]. Although these studies were carried out to investigate the impact of agitated filter drying condition on particles properties, the focus of their investigations were limited to the assessment of the drying rate and the change in particle size distribution on a microscopic scale.

One of the most challenging issues in the operation of an AFD is the formation of large agglomerates that may reach up to 10–15 mm in size, depending on the scale of the equipment. To date, there have been only two papers that investigated the occurrence of large agglomerates in an AFD operation and suggested strategies to counter the formation of these undesired lumps [7], [8]. am Ende et al. [8] assessed the risk of agglomeration and predicted the propensity of various API powders in forming undesired agglomerates through the use of mixer torque rheometer. The risk of agglomeration or attrition was quantified using the maximum torque and the moisture content where the maximum torque occurred. Through this method, various types of wash solvents were screened to select the best powder-solvent combination that produced the lowest agglomeration potential. The authors also suggested an extended blow-down or static drying to avoid agitation when the powder bed is above the “critical moisture content”. The “critical moisture content” was defined as the moisture content where the wet powders will have a maximum agglomeration tendency, and agitation will lead to severe agglomeration [1], [2], [7], [8]. Similarly, Birch and Marziano [7] also investigated the solubility and rheological aspect of the agglomeration behaviour of cohesive powders in a lab scale filter dryer. The occurrence of agglomeration was reduced to negligible levels by reducing the initial moisture content before drying using a blow-down step and avoiding agitation at its critical moisture content, coupled with the use of solvent with lower particle solubility. However, the reduction of moisture content using the blow-down step in order to reduce the propensity to agglomerate may not be possible for all APIs as many APIs have a minimum achievable moisture content where beyond this moisture content, heating is required to drive the remaining bound liquid out from the particles. The minimum achievable moisture content for each powder depends on factors such as particle size distribution, shape and the amount of hydrophilic surface functional group. These contribute to the differences in surface energetics between powder types which potentially creates differences in interaction and binding of water molecule to the crystal surfaces [16]. In addition, differences in the pore network distribution of powders also play a significant role where water may be physically trapped within the pore structure. Some powders may have a moisture content of 40 wt% even if a blow-down period of 24 h was applied. The studies to date [7], [8] so far have recommended agglomeration prevention methods that are specific to the powder type used. Whilst this is a good progress in a relatively under-studies area, more systematic and mechanistic approaches are needed to eliminate the agglomeration phenomena for all powder. There is a lack of understanding in the agglomeration formation mechanisms and the effectiveness of each mitigation method in reducing the lumps due to the unique “characteristics” exhibited by the different types of powder.

In addition, the studies had so far focused on the operation condition and the properties of the material and solvent itself. In practice, it is known that other external factors such as the presence of impurities and dripping of condensed vapour may play a significant role in the formation of agglomerates as well. The impact of impurities will not be covered in this work due to its complexity where different products may contains different types of impurities which ultimately depends on manufacturing method and how the raw material is obtained. On the other hand, the dripping of condensates is a common phenomenon in an AFD process which occurs primarily when the drying rate is high (Fig. 2). This may occur at high drying temperature or at the start of the drying when the moisture content of the powder bed is at its highest. The rapid evaporation of solvent at high drying rates will generate large amount of vapour which creates a highly saturated environment over the headspace of the dryer. The drying process in a typical AFD is carried out through heat sources from the heated agitator and a jacketed wall. However, an AFD may contain equipment parts that are not heated during the drying steps such as the spray bars used for cake washing, the sight glass for inspection, the ceiling of the AFD and the vapour exhaust (vacuum line). All these equipment parts will act as cold surfaces where the generated vapour is likely to condense and eventually drip back down into the powder bed, leading to the nucleation and growth of agglomerates. Nucleation occurs when liquid was brought into contact with the powder bed, leading to the formation of nuclei or primary agglomerates. The effect is more significant if the liquid is not distributed evenly throughout the powder which is a function of the powder-liquid combination as well as their mixing efficiency [17]. The possibility that agglomerates can be generated as a result of condensate drips will be investigated in this work.

The overall aim of this work is to understand the agglomeration behaviour of different powders in an agitated filter dryer. First, the effect of operating conditions such as temperature, agitation speed, blow-down period and height of the powder bed on the formation of agglomerates will be investigated. Second, it was hypothesised that the dripping of condensates forming on the cold surface within the vessel may induce nucleation or growth of the agglomerates. This will be investigated by studying the colour pattern formed in the agglomerates using dye solutions as tracers to mimic the condensates as well as the moisture profile of the agglomerates. Finally, a binary solvent mixture was used to re-slurry the powder to investigate the role of solubility and surface tension on the formation of agglomerates. It should be noted that different powder types exhibit different agglomeration tendencies which can be contributed by a number of factors including: physical properties such as particle shape, surface roughness and solubility as well as the surface chemistry of the exposed crystal faces. Hence, a direct comparison of agglomeration tendencies between powders may not be straightforward. This paper will demonstrate how a comprehensive understanding of agglomeration behaviour and its key properties is required before an effective mitigation method can be carried out due to the differences in the underlying agglomeration mechanism for each powder.

Section snippets

Materials

The materials used for this study comprised of two food grade model powders and an API intermediate. The model powders used were calcium carbonate (CaCO3) from Melbourne Food Depot, Australia and sodium bicarbonate (NaHCO3) from McKenzies, Australia. The API intermediate which exhibits severe agglomeration behaviour was provided by GlaxoSmithKline. The particle size distribution for CaCO3, NaHCO3 and the API intermediate was measured using Malvern Mastersizer and is shown in Fig. 3. The

Agglomerates formation

In this section, the effect of operating conditions on NaHCO3 and CaCO3 was investigated using a DoE approach to understand the agglomeration behaviour of free-flowing and cohesive powder respectively. The agglomeration behaviours of these powders were compared to the API intermediate which exhibits high agglomeration tendency.

The surface plot in Fig. 4 depicts the effect of varying the operating conditions on the formation of lumps. An increase in temperature was found to promote the formation

Conclusion

This study focuses on understanding the impact of process conditions (drying temperature, agitation speed, blow-down period and fill volume), dripping of condensate and the choice of wash solvent on the formation of agglomerates (> 2 mm). The agglomeration behaviour of two food grade model powders (NaHCO3 and CaCO3) and an API intermediate subjected to filter drying was investigated. It was shown that the minimization of agglomerates in the model powders can be achieved through proper selection

Acknowledgements

The authors would like to acknowledge GlaxoSmithKline and Monash University for providing financial support and resources for this study. Note that since September 2015, the Port Fairy facility where this study was conducted is now owned and operated by Sun. Pharmaceuticals.

References (30)

  • S. Murugesan et al.

    Design Of Filtration and Drying Operations

  • B.A. Perlmutter

    Application of pressure nutsche filtration and drying technology for toxic chemicals

    Process. Saf. Prog.

    (1999)
  • K. Tadeusz et al.

    Special Drying Techniques and Novel Dryers

  • M. Birch et al.

    Understanding and avoidance of agglomeration during drying processes: a case study

    Org. Process Res. Dev.

    (2013)
  • D. am Ende et al.

    Development and application of laboratory tools to predict particle properties upon scale-up in agitated filter-dryers

    org. process res. dev.

    (2013)
  • Cited by (17)

    • Environmental assessment of metal-organic framework DUT-4 synthesis and its application for siloxane removal

      2021, Journal of Environmental Chemical Engineering
      Citation Excerpt :

      Similar results were obtained for the hydrothermal synthesis (H2O, 72.8 dyn/cm), where particles of different sizes exceeding 3.25 µm were found. The formation of larger agglomerates during drying of the DUT-4(DMF) and DUT-4(W) was due to the capillary forces overcoming the repulsive forces between the particles; therefore, the particles came closer to each other [45,46]. In summary, according to the FTIR and NMR analysis, the spectra of the materials obtained by solvothermal and hydrothermal synthesis show the characteristic absorption bands of the functional groups of the organic linker and the chemical shifts related to the six-fold coordinated Al, respectively.

    • A Compact Device for the Integrated Filtration, Drying, and Mechanical Processing of Active Pharmaceutical Ingredients

      2020, Journal of Pharmaceutical Sciences
      Citation Excerpt :

      The isolated product typically comes with a large amount of dry lumps, thus requiring separate steps for delumping and particle size reduction. The degree of aggregation during the drying process is difficult to control, as it depends on a wide variety of mechanisms, including solvent composition, moisture content, and impurities.14,15 Consequently, a common approach is to deal with the aggregated material in separate delumping and particle size reduction steps after the drying process.

    View all citing articles on Scopus
    View full text