Improving liquid distribution by reducing dimensionless spray flux in wet granulation—A pharmaceutical manufacturing case study

Abstract

Controlling liquid distribution in a wet granulation process is critical to maintaining control of both nucleation and growth, as highly saturated patches created by uneven liquid distribution will have a much higher growth rate and may form large “balls”. The size of these balls depends on the granulator scale, but in full-scale pharmaceutical manufacturing these lumps are commonly 2–3 cm in diameter, but may even be larger. The presence of a significant quantity of “balls” frequently results in downstream problems in drying, milling, compression and final product attributes.

This paper presents a case study of attempting to improve the liquid distribution during manufacturing of an existing wet-granulated product with a long history of “balls”. The flowrate and spray area of the original nozzle were measured, and a simple estimation of the drop size was obtained by a high-speed photograph and image analysis. The powder surface velocity in a 400L Diosna was measured using a high-speed camera and a simple image analysis technique for several batches using different lots of the drug. Since the manufacturing process was validated and filed, the simplest process change that could be made to attempt to reduce the spray flux and improve liquid distribution was to find a new nozzle. A new nozzle was selected and implemented in a full-scale production batch and the results are compared with the original nozzle conditions. Reducing spray flux by changing the nozzle actually increased ball formation, contrary to what was expected. For products in the growth and/or induction regimes, increasing the efficiency of liquid distribution may mean that less total liquid is required to be added to achieve the same extent of granulation. Improving liquid distribution without also reducing the total volume will result in a shorter induction period and/or a higher growth rate and larger granule size. The study found that the major contribution to batch-to-batch variation in spray flux was the large variations in powder surface velocity for each batch, which is presumably caused by changes in the physical properties of each lot of drug. This has not been reported previously and has important implications for understanding the causes of variability in liquid distribution and granule/ball size in full-scale production of wet-granulated pharmaceutical products.

Introduction

Granulation is a processing technique commonly used in the pharmaceutical industry to creating porous, free-flowing granules from a mixture of dry powders. During wet granulation, all the dry powder ingredients – the drug and typically 3–8 other excipients – are mixed vigorously while a granulating fluid is added, which may be water or a volatile but non-toxic solvent, such as ethanol or isopropanol, or a mixture. These solvents can be used directly as a granulation liquid or as a delivery agent for a polymeric binder. Binders are a polymeric excipient used to ensure that the particles are still being held together after being dried and commonly come in several grades of varying molecular weights and physical properties.

The granulating liquid is typically sprayed onto to moving powder bed, provided that the fluid is not too viscous. This liquid will be gradually dispersed by wetting and capillary action combined with mechanical agitation from the impeller. The particles will then bond to each other either due to being immersed in a drop, or as a result of the distributed liquid films creating bridges between the particles. In either case, a distribution of nuclei granules will be created – this stage is called wetting and nucleation [1]. The granulation mechanism will then follow by consolidation and growth where the granules are colliding with each others, or with other powders, to form bigger granules [1]. The last stage is attrition and breakage. At this stage, the granules break due to impact, wear or compaction in the granulator [1].

In current industrial practice, the granule is not directly controlled. Instead, the manufacturing proceeds according to a set batch “recipe”, which commonly sets the total liquid quantity to be added, the liquid addition time (i.e. spray rate) and the impeller speed. There may also be an extra period of mixing after the end of the liquid addition step, which is often referred to as “wet massing”. During product development, the granulation recipe is set to provide a process which is as robust as possible. However, it is common on scale-up, or over time during the 10–50-year life of a product, for the process conditions to drift, due to cumulative small changes in equipment (nozzles, pumps, excipient suppliers and grades) and/or variations are the physio-chemical properties of the incoming drug. This latter problem has become more prevalent recently, as the need to reduce manufacturing costs often requires that alternate cheaper sources of a drug be used. It is impossible to completely specify all of the properties of a powder and eliminate all variation, even if all the required properties were known. Although rarely seen in development, all full-scale pharmaceutical manufacturing sites struggle with batch-to-batch variation in raw materials, unit operation performance and subsequently the product performance. Most manufacturing sites have one variable product, and these products often (but not always) have relatively high drug loads. Since there is no on-line, real time control of granule size during manufacture, operators often find that the batch contains three broad types of granules – fine particles (say <100 μm), small to medium granules (100 μm to 1 mm) and “lumps” or “balls”. The size of these balls depends on the granulator scale, but in full-scale pharmaceutical manufacturing these lumps are commonly 2–3 cm in diameter, and may be even be larger (see Fig. 1).

Lumps or balls are undesirable, and cause multiple downstream processing problems can arise. During pharmaceutical manufacturing of a wet-granulated product, the wet granules are dried in a fluidised bed and then passed through a size-selective mill. The drying time is proportional to granule size, and the large lumps are difficult to dry and make it difficult to accurately determine the average moisture content of the batch. Even when the moisture content is low enough to seem “dry” and the drying operation is ended, the large 0.5–1 in. balls often have dry exteriors surrounding a wet centre.

After drying, the granules are milled in a size-selective mill. A screen retains the granules in the milling chamber, where a high-speed impeller is rotating. Granules smaller than the screen hole size (around 0.5–1 mm) pass quickly through with little comminution, while larger granules are held in the milling chamber and subjected to impact and shearing from the impeller blades, until they are small enough to pass through the screen holes and exit the milling chamber. If there are too many large balls which need substantial milling, the milling chamber can become clogged or blocked, and heat build-up can adversely affect the product stability. The large balls experience substantial attrition, which generates new fine particles in the product, effectively “undoing” some of the granulation process. The milled granule flow and ability to form strong tablets are both adversely affected by fine particles. In addition, if the balls contain a wet centre, the moisture is released during milling and can smear the inside of the mill, or re-humidify the entire batch.

Once a batch is in process, there is currently no way to correct or adjust the processing conditions to fix these problems. The only place where adjustments can be made is during tablet compression, where the fine and/or damp powder results in variable tablet weights due to inconsistent flow and variations in tablet hardness and thickness data due to the small particle size or moist patches. Often these issues can be overcome by adjusting the compression force or slowing down the press, but at the very least this represents reduced manufacturing efficiency due to lost time and product while troubleshooting to find a new set of successful tabletting conditions. Ultimately, the failure to reach the specified tablet hardness or thickness may result in batch rejection, which is extremely expensive.

Clearly, it is desirable to minimise the presence of large lumps or balls in a wet-granulated product. There are two main ways that these large balls are thought to form during granulation:

• Rapid formation and growth of granules due to the presence of highly saturated regions of powder within the granulator (i.e. wet patches).

• Rapid uncontrolled coalescence of granules, often in the induction growth regime [2]. At the onset of growth, the granule porosity is at a minimum and the granule saturation reaches a maximum, producing surface-wet granules. The tendency for induction behaviour is set by the formulation and the process conditions affecting granule consolidation.

This paper summarises an industrial case study focusing on the first mechanism, where wet patches form large, wet nuclei which then grow rapidly due to their higher saturation. Controlling liquid distribution in a wet granulation process is beneficial to maintaining control of both nucleation and growth, as highly saturated patches created by uneven liquid distribution will have a much higher growth rate and form large balls. After a brief review of the literature on spray flux and liquid distribution, we present a real industrial case study calculating the dimensionless spray flux and using this information to improve the liquid distribution by selecting a new nozzle. The case study was performed on an existing wet-granulated product with a long history of “balls” and unwanted variation in granulation and tabletting performance.