Michael Thompson is a professor at McMaster University.

Michael Thompson

In less than a decade, foamed systems for continuous twin-screw granulation have leaped from the theoretical confines of academia onto the mainstage of industry relevance. Once viewed as a fringe technology with a lot of potential, the method of adding a foam binder during twin-screw granulation is now a commercially viable way to improve robustness.

	Foam granulation introduces binding liquid into a powder bed without oversaturation; it is a tool to solve problems associated with injection wetting during twin-screw granulation. Using foam, formulators can achieve more uniformly wetted powders and increase the overall lubricity throughout the process, which minimizes both wear and tear on machinery and dissipative heating of the product on continuous manufacturing production lines.

	Continuous granulation-of which twin-screw is the most prevalent type-can streamline overall manufacturing processes by reducing costs and increasing output while still maintaining consistent product quality. Generally speaking, continuous granulation depends on a host of interconnected variables, including process temperature, screw speed and configuration, powder feed rate, and the introduction of moisture into the powder bed. If any one of those variables is improperly adjusted, the process is compromised. For example, using a liquid stream to wet the powder bed (i.e., injection wetting) can lead to issues such as uneven moisture, temperature spikes, and inconsistent component distribution.

	Twin-screw granulation sets itself apart from other continuous granulation methods because of its minimal operating space. It also enhances flowability and compressibility, effectively increasing the efficiency of downstream processes, such as tablet compression or capsule filling.

	As with any process, twin-screw granulation isn’t perfect. Problems with oversaturation can arise because of the confined internal operating space, which makes aerosolizing the liquid feed almost impossible. Oversaturation of the powder bed during wetting can ruin granulation, especially considering the costly demands of downstream milling and drying steps that must compensate for inadequacies in the process itself. Liquid injection streams can also lead to segregation of APIs, with substantial heterogeneity resulting in larger granules that contain more liquid due to the dominance of layering growth in a shear flow field.

	A foam delivery system has been discovered to solve oversaturation problems associated with introducing liquids into the powder bed, but this method only works if the foam is treated as a collapsing solid when designing the twin-screw granulation process (1). By wetting the powder bed differently, many of the challenges involved in designing a twin-screw granulation process can be overcome.

	The advantages of the novel foam delivery system are primarily driven by the high spread/soak ratio enabled by foam, which forces liquid to move slowly and evenly along the boundaries of gas bubbles, extending its penetration time to match the time scale of bulk mixing that occurs within the extruder. The result is improved liquid distribution in the powder bed for the short residence time distribution of this type of granulator. 

	Foam granulation allows for a uniform bed wetting, which prevents product temperatures from rising because of improved liquid distribution and the lubricious nature of foam. By extension, foam can protect thermally sensitive products, decrease variability, and create a more stable process.

	Unlike liquid injection methods, API concentrations remain constant across the process and in the final particles when foam delivery is used, preventing segregation even when powders are highly hydrophobic. Particle size does not appear to impact API distribution as it does with liquid injection, and binder distribution (as monitored by tracer) is significantly more uniform (2). In traditional granulation, the imprecise nature of liquid injection over a deep bed of powder results in a small portion of solids getting mixed with the blade. But by implementing foam delivery within a twin-screw granulation process, the uniform wetting of a much shallower powder bed results in nearly all solids getting mixed. This combination of processes could lead formulators to novel powdered drugs that couldn’t be produced otherwise.

	Because of the highly active exploration into twin-screw granulation in both academia and industry, process improvements that were previously unrealistic are now available to almost any researcher or manufacturer. These improvements are well within reach, too, as foam delivery systems can be set up at low cost using mechanical components that are already commercially available from various vendors. 

	To further understand the mechanism and benefits of foam granulation, consider an experiment by DuPont researchers involving two powder beds: one of pure lactose monohydrate and one containing a 50:50 blend of lactose and microcrystalline cellulose (MCC) (3). Two liquid introduction methods-droplets (analogous to liquid injection) and foam-were used. The team looked at the liquid penetration time for each of the four combinations, finding it takes approximately 100 times longer for foam to saturate a bed of powder made up of either lactose or MCC versus simply dropping liquid onto a similar bed. 

	Researchers began the study by sieving each formulation through a 300-µm screen. They then poured a sample of the powder (roughly 100 g) through a funnel suspended 5 cm above a glass petri dish measuring 10 cm in diameter and 1.5 cm deep. The team prepared the powder bed by evenly spreading powder around the perimeter of the dish to minimize differences in porosity between repeats. Variations in porosity using this method were observed-roughly 1.89% relative standard deviation. After preparation was complete, they added differing masses of either foam or liquid lactose monohydrate on top of the powder bed. 

	Researchers monitored the dispensed liquid until it fully penetrated the bed, then left the binder-saturated powder bed for 48 hours to air dry, before they removed the resulting granule nuclei. They defined the specific penetration time (t

) as the elapsed time until the binder had fully penetrated and normalized with respect to the binder mass, whereas the nucleation ratio (Km) referred to the mass of the resulting granule nuclei relative to the binder mass. This penetration study corresponds to the method devised by Hapgood and coworkers (4, 5).

	The passage of liquid through the bed consistently took longer for pure lactose compared to the 50% MCC formulation because of its lower affinity for water. That difference was only statistically significant when looking at foam delivery. The difference between lactose and MCC was consistent with results reported in an earlier study (6). As previously found by Hapgood (4, 5), the smaller the dispensed binder mass on top of the powder, the longer the time needed to fully penetrate by either method of wetting to the state of powder saturation as it affects liquid mobility. The difference in the specific penetration time between foam versus droplet was significant-in fact, foam took almost two orders of magnitude longer to fully penetrate the powder than liquid. This difference is attributed to the fact that liquids drain from a semi-rigid foam more slowly, following a methodical path along the plateau borders and film walls of gas bubbles-rather than allowing all of its retained liquid to be immediately available like a droplet (7).

	The researchers also found that a foam binder addition can prevent temperatures from rising because of improved liquid distribution and the lubricious nature of foam. 

	Commercial acceptance of foam granulation

	Although continuous processing offers advantages, there is resistance in an industry that has relied on time-tested batch production processes. The pharmaceutical industry is seeing a broader range of new products in continuous hot melt extrusion using a twin-screw extruder rather than with any form of wet granulation (with a twin-screw or other technology). Research in twin-screw granulation, however, is finding some interesting developments. For example, the process naturally makes particle-size distribution that is uniquely bimodal, as opposed to the unimodal distribution made by batch processes. 

	Moving away from a traditional wet granulation process also opens the door for formulators to start using more hydrophobic APIs, ensuring that binders are properly distributed to accommodate a new range of difficult-to-process formulations. As more APIs become mixable, new products become available. Additionally, a reduction in water usage brings sustainability benefits to the table.

	Formulators can implement foam granulation to save time compared to batch processes, because many of the steps associated with batch processing are removed from the workstream-including weighing of materials, charging of the granulator, changeovers between batches, cleaning between batches and transfer-associated issues. Furthermore, by eliminating many of the common challenges associated with proper design and implementation of liquid injection twin-screw granulation-such as proper liquid distribution and preventing in-process component segregation, for example-it is possible to arrive at the final process and formulation more quickly than before.

	While in the past continuous foam granulation may have been seen as a fringe technique to hypothetically solve process-surging issues that relate to poor powder-wetting by conventional, liquid-addition methods, understanding the high spreading tendency of foam granulation now makes it a valuable technique for commercial product development. 

38 (7) 771–784 (2012).
	2. H. Li, M.R. Thompson, and K.P. O’Donnell, 

, 496 (1) 3-11 (2015).
	3. Y. Li et al., 

 61 (3) 780-791 (2015).
	4. T. Nguyen, W. Shen, and K. Hapgood, 

64 (24) 5210-5221 (2009).
	5. M. Tan et al., 

64 (12) 2826-2836 (2009).
	6. K. Rocca et al., 

41 (1) 35-42 (2015).
	7. C. Keary and P. Sheskey, 

	Kevin O’Donnell is R&D manager at DuPont Nutrition & Biosciences; Michael Thompson is Professor of Chemical Engineering, McMaster University.


	Vol. 44, No. 7
	July 2020
	Pages: 30–33

	When referring to this article, please cite it as K. O’Donnell and M. Thompson, "Innovating Continuous Twin-Screw Granulation with Foam Delivery," 

Articles

Using a specific addition method to wet the powder bed in twin-screw granulation using a foam binder improves robustness for continuous solid-dosage drug product manufacturing.

Innovating Continuous Twin-Screw Granulation with Foam Delivery

As a result of changing philosophies towards continuous manufacturing, new equipment is being introduced into pharmaceutical production facilities. The twin-screw extruder is an example of such equipment for use in wet granulation. The authors review developments in wet granulation using a twin-screw extruder; lay out the issues with wetting in this machine; and introduce a novel technique, foam granulation, that uses the twin-screw extruder to fully satisfy the unique needs of granulation.

The twin-screw extruder provides highly consistent granulates due to its continuous operation and closely confined flow path, which requires that all particles experience a similar shear history. The intensive mixing of the twin-screw extruder allows lower optimal liquid concentration for granulation while producing denser granules for both placebo formulations and highly dosed drugs in comparison to a high-shear batch mixer (1, 2). As a result, drying and milling operations may be significantly reduced with use of this machinery in solid oral-dosage production. 

The binding liquid in wet granulation has a profound influence on product granule properties (3–5) and affects the friction between conveyed powders and the barrel wall inside the extruder, which affects power consumption and the exiting temperature of granules (2, 6). There are crucial issues to be solved in regards to introducing liquids into this type of machinery to obtain rapid and uniform wetting of excipients so that the process exhibits stability in operation, boundaries become immediately lubricated to reduce equipment wear and granule heating, and high quality granulates are obtained.  

A common variant of extruder used for granulation is the fully intermeshing, co-rotating twin-screw extruder (7). Differences between vendors are largely based on the available internal volume of the machine (often described by the ratio of the outer diameter to inner diameter of screw elements) as well as the screw diameter, both of which can significantly affect granulate properties in both granule size and intragranular porosity (8). The machine is highly modular, making it a flexible platform for continuous manufacturing of different products during its lifetime of service to a company. The intermeshing region between the two screws creates a self-wiping action that minimizes material accumulation within the machine but also provides a complex flow path for powders to mix and consolidate. For wet granulation, the die end of the extruder is generally open to collect granules without excessive consolidation. 

Wet granulation inside the co-rotating twin-screw extruder is a starve-fed process, meaning that the available internal volume of the machine is never completely filled with material during operation. This 

is important to extrusion because it minimizes dissipative heat build-up in conveyed drug formulations as it limits compression against the barrel wall, it decouples the parameters of output rate and screw speed to give formulators more control over their process, and it more readily allows the downstream addition of materials (solids, liquids, or gases) because the system is not pressurized except for small mixing regions. The zones of the screws that are starved experience dominant drag flow, in which powders are pushed downstream by the rotating flights of conveying-type elements. These screw elements have been found to contribute little to granule growth (5, 9). In fact, screw designs using only conveying elements show very poor distribution of the binding liquid within exiting solids (10). It is rare, however, that a screw design is completely comprised of conveying elements or that the entire length of the machine is ever fully starved. Significant granule growth requires the inclusion of pressure-driven mixing zones, which are necessarily fully filled as powders are squeezed through these sections. Kneading blocks and comb (i.e., chopping) elements are examples of mixers commonly used in sparing numbers along the screw length to produce granule growth along with minor attrition (5, 9, 11–13). 

 shows some typical designs for granulation. Keeping these mixing elements closer to the end of the extruder reduces attrition (4). 

Figure 1: Characteristic screw designs used in twin-screw extruders for wet granulation. Screws displayed correspond to literature referenced in this article: (a) ref. 11, (b) ref. 4, (c) ref. 6, and (d) ref 6. Powders enter at the extreme left-hand side and arrows indicate points of liquid addition. (ALL FIGURES ARE COURTESY OF THE AUTHORS.) 

Powder flow rate is one of the most significant parameters influencing the extent of granule growth, with higher outputs producing larger granules. The effect is caused by the higher volumes of powder that build up in front of pressure-driven mixing zones as flow rate increases, producing larger axial compressive forces on the particles present. In fact, it has been shown that the dispersion of binder within poorly wetted mass can be improved for granulation if the screw design and flow rate are adjusted to provide appropriate compressive forces (6). The influence of flow rate on granule growth, however, is not often seen in smaller extruders or highly starved processes (4). Increasing screw speed has less influence on granule size but generally increases the number of chopping events provided by mixing zones to reduce the occurrence of oversized particles (4, 6, 9). For a fixed flow rate, increasing the screw speed will reduce the volume of powder that fills the conveying screw elements, resulting in lower power consumption by the process.  

Among the published studies for wet granulation, a crucial point that is rarely mentioned, yet widely known to the pharmaceutical industry, is the difficulty of uniformly wetting a formulation in an extruder. The problem arises due to the earlier mentioned closely confined space inside the extruder, which results in the liquid injection port being in immediate proximity to the powder flow. This confinement prevents atomization of the binder solution into micro-sized droplets prior to contacting the powder solids, as is done in high-shear batch mixers. As a result, regions of the powder become oversaturated while others remain virtually dry. This issue was highlighted in the industrial-oriented article by Shah, who reported process surging, though motor overload events are also common (11). Shah demonstrated several strategies related to screw design and the sequential addition of smaller liquid quantities into the process as means to minimize surging occurrences. Such changes greatly increase the complexity of operating the extruder and do not eliminate the root cause of the problem. Alternatively, a new solution called foam granulation uses the unique behavior of aqueous foam to cause rapid spreading of the binding liquid over a large area of the powder during wetting.   

The aqueous foamed binder used in foam granulation is comprised of a high volume of gas dispersed within a liquid containing foamable excipients, thus forming an unstable, semi-rigid structure. Effective excipients for pharmaceutical granulation are cellulose-ether species that promote high foaming activity and act as binders in the process. Many approved nonionic, polymeric excipients are also suitable foaming agents. The foam liquid may include additives (e.g., polymeric species for binding or coating and particles, such as APIs, glidants, and disintegrant aids) as long as they do not interfere with its preparation. Semirigid foams characteristically exhibit closely packed bubbles or a polyhedral morphology depending on the gas-volume fraction although a minimum of 64% (v/v) gas is required for the foam to display some degree of rigidity. The volume fraction of gas present in foam is often referred to as its foam quality (FQ). For granulation, FQ is generally kept in a range of 75–95%. Foams that are too wet (< 75% FQ) lack adequate stability to spread well and often simply collapse on the surfaces of processing equipment. Very dry foams (> 95% FQ) occupy very large volumes of space (which complicates their addition into the confined process); exhibit very high inherent viscosities (as much as 10

  times that of its contained liquid); and more readily collapse in the presence of shear than wetter foams (14).   

Foam granulation was first introduced by Keary and Sheskey for high-shear batch mixing of pharmaceutical ingredients (15). This study demonstrated that less binding liquid was required and that the rate of foam addition could be much higher in comparison to spray wetting. The lower requirements of binding liquid were explained in the staticbed penetration studies of Tan 

., which looked at saturation characteristics of foam versus dropwise wetting with lactose and glass ballotini powders (16). These studies observed that more binder mass was absorbed by these powders by foam as opposed to dropwise addition at any given time, and as a result, granule nuclei were 40% larger. Foams of higher FQ were more slowly absorbed due to slower foam coarsening and slower drainage of its contained liquid into contacting powder (14, 17). Several studies of foam granulation for high-shear batch mixers have been reported (17–21). 

Figure 2: Typical equipment layout for foam granulation showing: (1) twin-screw extruder, (2) powder feed-port, (3) gravimetric feeder for powder excipients, (4) side stuffer, (5) mechanical foam-generator, and (6) foam feed-tube.   

Continuous foam granulation with a twin-screw extruder was introduced in a case study comparing the technique to the conventional liquid addition method (6). A successful methodology to metering such foam into the machine required recognizing its solid-like behavior and using approaches commonly employed for feeding bulk solids rather than liquids. An auxiliary unit, known as a side stuffer to the extrusion industry, was found suitable for feeding foam. The side stuffer is readily available commercially, and the physical setup and control software of most extruders can be configured to accommodate it. A typical extrusion setup with the side stuffer and foam generator is shown in 

. The side stuffer is a miniature, twin-screw auger that mounts to the side of the main extruder and conveys materials into a specified zone of the process. Due to the drag-flow action of the rotating screws in the side feeder, foam is forced into the passing formulation within the main extruder and partially collapses upon contact, while the remaining foam forms a layer between the powder and extruder barrel. 

 highlights the conceptual differences between liquid injection and foam addition from a cross-cut view of the extruder. The mechanism of foam wetting inside the extruder is still under study. A two-stage model proposed in a recent publication was based on how foams prepared from liquids of different viscosities and having different FQ collapsed and drained under different shear conditions as well as how they affected granule properties from the extruder (14). A pressure-driven wetting stage is thought to occur at the point of entry where the foam enters the process, with stiffer foams showing greater resistance to immediately collapsing upon contacting the non-wetted formulation. The remaining, uncollapsed foam pushes the powder aside to form a layer above. The subsequent shear-driven wetting stage appears governed by the response of foam to shear; layers of stiffer foam collapse more readily under mechanical shear to wet the powder beneath while wetter foams show greater resistant collapse under mechanical shear by establishing more stable morphologies comprised of smaller bubbles. 

Figure 3: A cross-sectional view of the twin-screw extruder partially filled with powder excipients (in blue) shows the differences in configuration used for (a) directly injecting liquids versus (b) introducing foamed binder into the granulation process.  

From a practical, operational standpoint, the method of foam granulation: 	  

Avoids process surging in a twin-screw extruder due to the high coverage area of foam over the powder during wetting 

Simplifies process start-up because full operation rates can be immediately achieved, as opposed to liquid injection, in which powder flow must be slowly increased to prevent motor overload  

	Reduces machine wear, as indicated by the extruder no longer experiencing the loud knocking noises indicative of screw deflection caused by uneven powder flow.

 These observations are thought to be related to the two-stage wetting mechanism previously described, which causes the powder to become immediately isolated from the barrel wall by a layer of foam, at least until it is well wetted. The powder in this case is steadily saturated with the binder over a much larger area of contact than in direct liquid addition, which minimizes the binder's local concentration in the porous matter. The lubricating feature of foam granulation, in which the foam layer isolates the powders from the barrel wall until uniformly wetted, is an important point to be stressed for extrusion processing. The lubricity of conveyed solids affects both power consumption by the machinery as well as the exiting temperature of granules (2). 

Comparing foam and conventional wet granulation

A study at pilot-scale flow rates (20-40 kg/h) compared foamed-binder addition and direct liquid-injection on granulation (6). A methylcellulose binder (Dow, Methocel A15 PLV) was used at two concentrations, 6% and 11% (w/w), relative to a-lactose monohydrate powder. Two screws were tested in the work to produce differing axial compression characteristics (which was mentioned previously as an important factor for granule growth inside the extruder) with changing flow rate: one with a single pair of mixing elements producing lower axial compression (LAC) and a second with two pairs in series to provide a more restrictive flow path and higher axial compression (HAC). Notable differences between the two methods of granulation are summarized in 

. The granule properties from the study showed that comparable sizes and intragranular porosity were achieved by either method, provided appropriate conditions were used. The reduced requirement for liquid in the process was a comparable finding to that found with high-shear batch mixers (15). 

Table I: Differences between foamed and liquid binder addition on granulation. 

Wet granulation in twin-screw extrusion machinery has several key advantages over conventional methods, but to advance in acceptance for GMP production, its operations need to be better understood and challenges regarding process stability need to be solved. Continuous foam granulation is a new, robust technique that solves the process- surging issues that relate to poor powder wetting by conventional, liquid-addition methods. The high spreading tendency of foam in granulation, versus the immediate soaking nature of liquids, produces more uniformly wetted powders and increases the overall lubricity of the process, which benefits wear behavior of the machine and minimizes dissipative heating of the product. With comparable particle properties to conventional wet granulation, foam granulation gives formulators greater flexibility in achieving production goals.   

1. E.I. Keleb, A. Vermeire, C. Vervaet, and J.P. Remon, 

2. L. Tan, A.J. Carella, Y. Ren, and J.B. Lo, 

3. E.I. Keleb, A. Vermeire, C. Vervaet, and J.P. Remon, 

4. E.I. Keleb, A. Vermeire, C. Vervaet, and J.P. Remon, 

6. M.R. Thompson, S. Weatherley, R.N. Pukadyil, and P.J. Sheskey, 

7. W. Thiele, "Twin-Screw Extrusion and Screw Design," in 

, I. Ghebre-Sellassie and C. Martin, Eds. (Marcel Dekker, New York, NY, 2003), pp. 69-98. 

DOI: 10.1016/j.powtec.2012.05.045, May 29, 2012.  

12. B. Van Melkebeke, C. Vervaet, and J.P. Remon, 

13. N.O. Lindberg, C. Tufvesson, P. Holm, and L. Olbjer, 

14. M.R. Thompson, B. Mu, and P.J. Sheskey, 

16. M.X.L. Tan, L.S. Wong, K.H. Lum, and K.P. Hapgood, 

18. S.L. Cantor, S. Kothari, and O.M.Y. Koo, 

19. P. Sheskey, C. Keary, D. Clark, and K. Balwinski, 

is an associate professor in the Department of Chemical Engineering at McMaster University, Hamilton, Ontario, L8S 4L7, Canada, Tel: 905.525.9140, 

 is a principal research scientist at Larkin Laboratory, The Dow Chemical Company, Midland, Michigan.  

*To whom all correspondence should be addressed. 

Submitted: Sept. 6, 2012. Accepted: Sept. 28, 2012. 

The authors review developments in wet granulation using a twin-screw extruder.