A series of studies on dry granulation using MgSt as a lubricant have been conducted (He et al., 2007; Miguélez-Morań et al., 2008; von Eggelkraut-Gottanka et al., 2002). von Egg von Eggelkraut-Gottanka et al. (2002) used a dry granulator to control the roller gap and pressure for compacting two batches of different traditional Chinese medicine dry extracts, and employed multiple linear stepwise regression analysis to investigate the effects of process parameters and magnesium stearate dosage. It was reported that the disintegration time of tablets increased with the increase in MgSt concentration due to the increased hydrophobicity. They also suggested that the binding of MgSt to particles (inside the tablets) minimized the increase in disintegration time while retaining its function as a lubricant.
He et al. (2007) performed dry granulation on MCC (Avicel PH 102, 44–75 μm) containing 0.5% (w/w) MgSt. The mechanism of the loss of repeated processability of rolled powders, especially after the addition of MgSt, was evaluated through Heckel analysis, measurement of tablet tensile strength, and dynamic indentation. They concluded that work hardening occurred during the process, and over-lubrication due to the presence of MgSt appeared to be the main reason for the decrease in the mechanical strength of tablets. Miguélez-Morań et al. (2008) studied the dry granulation of MCC (Avicel PH 102) under three conditions: (1) unlubricated; (2) lubricated roller surfaces; (3) bulk lubricated powders. The results showed that when the powders were internally lubricated with MgSt, the feeding was the most uniform, the resulting flakes had the most uniform density, and the maximum pressure reduction could be observed during compaction. Their work clearly demonstrated that MgSt has an impact on the dry granulation of MCC.
Despite existing studies, the relationship between compaction performance and downstream processes such as sizing and tableting, as well as the lubrication mechanism, has not yet been established. In particular, the influence of lubricants on sizing behavior, and the properties of granules and tablets, are not well understood. These are precisely the focuses of the current research, in which MCC and DCPD are selected as feed powders. Both are commonly used pharmaceutical excipients but have unique particle sizes, surface morphologies, and sensitivities to MgSt lubrication. It has been reported that MgSt-lubricated DCPD is very insensitive to mixing conditions (Vromans et al., 1988), which is significantly different from MCC (Zuurman et al., 1999).
MgSt is a white, odorless flaky powder (see Figure 1). Calipharm Grade D DCPD (Rhodia, France) is a friable crystalline powder with shale-like particles (see Figure 2a). MCC (Avicel grade PH 102, FMC Biopolymer, USA) (see Figure 2b) is a crystalline powder (crystallinity > 78%) with acicular particles, exhibiting greater plastic deformation compared to the relatively brittle DCPD. The true densities of MCC and DCPD were measured using a helium pycnometer (AccuPyc II 1340, Micromeritics, USA) as 1569 kg/m³ and 2582 kg/m³, respectively. Using a particle size analyzer (Model Helos, SympaTec, Germany), the average particle sizes of the two materials were measured as 96.3 μm and 8.1 μm, respectively.
Fig. 1. Scanning electron micrograph of magnesium stearate.
Fig. 2. Scanning electron micrographs of DCPD and MCC powders with andwithout MgSt. (a) Unlubricated DCPD; (b) unlubricated MCC; (c) bulk lubricatedDCPD (0.75% w/w MgSt); (d) Bulk lubricated MCC (0.75% w/w MgSt); (e) bulklubricated DCPD after ring shear cell tests; (f) lubricated MCC after ringshear cell tests.
Different dosages of MgSt (0.15–1.5% w/w) were mixed with these two powders respectively using a double-cone mixer. Preliminary studies (not described here) showed that the friction and flow properties of the powders did not change when the mixing time exceeded 5 minutes; therefore, the mixing time for all tests in this report was set to 5 minutes. The surface morphologies of the powders lubricated with 0.75% (w/w) MgSt are shown in Figures 2c and 2d.
A RST-XS ring shear cell tester (Dietmar Schulze, Germany) was used to measure the effective internal friction angles of the feed powders and sized granules, with a normal stress range of 4–10 kPa. This instrument was also used to measure the wall friction angles against a smooth stainless steel plate (surface roughness Ra ~0.3 μm), within a normal stress range of 1.1–20 kPa. The morphologies of the powders before and after wall friction measurement were obtained using a scanning electron microscope (6060, JEOL, Japan), as shown in Figures 2e and 2f.
An experimental dry granulator developed by the University of Birmingham was used to compact the powders (Bindhumadhavan et al., 2005; Miguélez-Morań et al., 2008; Patel et al., 2010). It consists of two stainless steel rollers with a radius of 100 mm and a width of 46 mm, using gravity feeding. One of the feeding methods involves manually filling an initial fixed volume of powder into a hopper with a rectangular cross-section, and the excess powder was gently leveled off.
In the current study, the minimum roller gap (S) and roller speed (u) were fixed at 1.0 mm and 1 rpm, respectively. The angle θ was measured at the minimum roller gap, and the corresponding radial roller pressure (p) was measured by a piezoelectric pressure sensor (PCB 105C33, Techni-Measure, Studley, UK). This sensor was installed at the center of one roller to obtain the pressure distribution of the rollers, so as to study the effects of bulk lubrication and wall lubrication. In the case of wall lubrication, the surfaces of the metal rollers were lubricated with MgSt ethanol suspensions at concentrations of 0.25% and 1%, respectively.
The dimensions of the flakes (e.g., length, thickness, and width) were measured using a digital vernier caliper (Mitutoyo, Hampshire, UK) to calculate their volume, and thus their bulk density. The fracture energy of the flakes was measured using a universal mechanical testing machine (Instron, High Wycombe, UK) in a three-point bending configuration. The force-displacement data were integrated to determine the total work of fracture. The fracture energy was calculated as the ratio of the work to the fracture surface area.
The flakes were cut into segments of a specific size (approximately 22×22 mm) to reduce the impact of differences in shape and size. A vibrating sizing sieve (Figure 3, Coeply, AR 401) with a sieve mesh size of 630 μm was used, and the sizing speed was set to 200 rpm. A computerized balance was used to measure the granule throughput as a function of time.
The sized granules were compressed into tablets using a universal testing machine (Z030, Zwick Roell, Germany) with a stainless steel die featuring an inner bore diameter of 13.0 mm (Specac, UK). The tableting speed was 0.5 mm/s, which is comparable to the effective uniaxial component (i.e., horizontal speed) of the dry granulator. The filling amounts of MCC and DCPD powder granules in the die were 0.8 g and 1.1 g, respectively. Correspondingly, the initial powder heights of MCC and DCPD under a pre-compaction force of 5 N were 15.05 mm and 9.43 mm, respectively. Uniaxial strain compression was then applied at ratios of 0.66 and 0.53 to obtain tablets with similar thicknesses (e.g., 5 mm). The compression coefficients were determined from the stress-strain relationships in uniaxial compression using the same multivariate fitting method proposed by Patel et al. (2010).
Three scenarios of uniaxial compression for the feed powders were considered: (1) unlubricated; (2) powder-lubricated; (3) die inner wall-lubricated. Once the tablets were ejected, their dimensions (e.g., diameter and thickness) and weight were measured to calculate their bulk density. A universal testing instrument was then used to perform radial compression tests on the tablets to determine their tensile strength.
Figure 4 shows the wall tangential stress as a function of normal stress for feed powders mixed with different dosages of MgSt, where the gradient equals the friction coefficient. In the case of unlubricated powders, the gradients for DCPD and MCC were ~0.5 and ~0.9, respectively; when 0.75% (w/w) MgSt was used for bulk lubrication, the gradient for DCPD decreased to ~0.1, while the gradient for MCC remained unaffected. The effect of lubricant dosage on wall friction was investigated in terms of the wall friction angle (Фw) and the corresponding effective friction angle (Фe), as shown in Figure 5. The Фe of unlubricated DCPD was only slightly higher than that of MCC and did not decrease with increasing MgSt dosage, unlike MCC.
As shown in Figure 6, the values of Фe and Фw are functions of MgSt dosage in the sized granules. Only slight differences in Фe values were observed between the granules (Figure 6) and the powders (Figure 5). The wall friction of unlubricated MCC granules was greater than that of the feed powders, and lubrication reduced the wall friction. For DCPD, the effect of MgSt on the wall friction of granules was insignificant.
Fig. 4. Variation of the wall shear stress with normal stress for (a)DCPD and (b) MCC, with various amounts of MgSt in bulk
Fig. 5. Frictional angles of MCC and DCPD as a function of amount ofMgSt.
Fig. 6. Frictional angles of MCC and DCPD granules as a function ofamount of MgSt.A simplified model of dry granulation was developed by Johanson (1965), which divides the space between the two counter-rotating rollers into distinct zones: the slip zone, the nip zone, and the release zone. The slip zone is the area where powder enters the rollers; the powder slips along the roller surface, rearranges itself in this zone, and is only subjected to minimal roller pressure. The position where the powder flow velocity equals the roller speed is defined as the boundary of the nip zone. In this zone, the powder is dragged to the minimum gap and then compacted into flakes under increased pressure—powder densification mainly occurs in this zone. After passing through the minimum gap, the compacted flakes enter the release zone, where elastic recovery of the flakes takes place.
The maximum pressure and nip angle of the dry granulator are two key parameters determining its performance. The typical pressure distribution measured in the current work is shown in Figure 7, with a maximum pressure of ~100 MPa and a nip angle of ~8°.
Table 2 presents the solid fraction and fracture energy of the flakes. The solid fraction was used to compare the compactness of the tablets. Bulk lubrication significantly reduced the solid fraction and fracture energy of MCC, while wall lubrication caused a much smaller reduction. Under the same roller compaction conditions (i.e., roller gap and roller speed), the solid fraction of DCPD flakes was lower than that of MCC flakes, and both wall lubrication and bulk lubrication reduced the solid fraction of DCPD flakes. However, DCPD flakes were too brittle to measure their fracture energy.
Table 3 shows the solid fraction and tensile strength of tablets made from feed powders and granules with and without lubricants. The solid fraction of each powder and its corresponding granules was approximately constant, which is consistent with the experimental procedure. It can be observed that during lubrication, the solid fraction of MCC powders and granules was not affected; however, the solid fraction of granules was lower than that of powders, which is attributed to the loss of flakiness. Meanwhile, the tensile strength of the tablets decreased significantly with the increase in MgSt dosage. In contrast, the solid fraction of DCPD was not affected by lubricants, and the values for powders and granules were comparable; the tensile strength slightly increased with the addition of MgSt.
The wall friction coefficient of unlubricated DCPD was approximately 5 times higher than that of MCC (see Figure 5), which is because the interfacial shear strength of organic materials is lower than that of inorganic materials. Boundary lubricants provide a weak interfacial layer, with a friction coefficient typically of 0.1 (Bowden and Tabor, 1950)—this represents the generally achievable minimum value. This is similar to the measured value of unlubricated MCC; therefore, boundary lubrication is ineffective in reducing friction for MCC. Organic polymers are usually difficult to boundary-lubricate, as they typically have similar interfacial shear strength to organic boundary lubricants. However, the friction coefficient of unlubricated DCPD was 0.5; thus, the application of MgSt could effectively reduce its friction coefficient to the observed minimum of 0.1. Obviously, a minimum MgSt dosage is required to form a uniform and stable surface layer.
The wall friction coefficient of unlubricated MCC increased after granulation, while that of DCPD decreased (see Figure 6). This may be attributed to the increase in particle size and changes in surface morphology. For granules produced from bulk-lubricated feed powders, boundary lubrication was only effective for MCC. The wall friction of DCPD granules practically did not change after the addition of MgSt, which may be due to at least two reasons. First, during sizing, some DCPD particles (with a small particle size of 8 μm) were lost; due to their large surface area-to-volume ratio, this led to a disproportionate loss of adsorbed MgSt. Second, even if the surface of DCPD particles was well-lubricated initially, the fragmentation of these relatively brittle particles would expose unlubricated internal surfaces (De Boer et al., 1978).
The effective internal friction angles of the two unlubricated powders were essentially the same, despite significant differences in their wall friction properties (see Figure 5). However, other factors such as particle shape and particle size distribution may be important for explaining this observation. When DCPD was bulk-lubricated, Фe did not decrease. This indicates that shear-induced particle fragmentation exposes fresh, unlubricated surfaces, which is consistent with the brittle nature of DCPD. Clearly, this mechanism does not apply to MCC. Wall lubrication caused a slight reduction in the wall friction coefficient, which necessarily explains the relatively large decrease in Фe—there is a considerable amplification effect from individual particles to the bulk.
In the case of shear cell data, it is impossible to directly compare the absolute values of maximum pressure and nip angle based solely on relative frictional properties. Other factors, particularly compressibility, need to be considered. Future work will involve verifying the accuracy of theoretical models (e.g., Johanson, 1965) for predicting the maximum roller pressure and nip angle based on the properties of feed powders. However, the measurements of wall friction and internal friction can provide some qualitative explanations for the changes caused by lubrication.
For MCC, lubricating the rollers affected neither the nip angle (Figure 10) nor the maximum roller pressure (Figure 9), which is consistent with the insensitivity of friction to wall lubrication. However, for bulk lubrication, both parameters decreased with the increase in MgSt dosage—this is reflected in the reduction of flake density and strength. As mentioned earlier, boundary lubrication of MCC is difficult. Nevertheless, bulk lubrication reduces internal friction and improves the overall flowability of the powder, thereby decreasing the nip angle and maximum pressure.
In the case of DCPD, unlike MCC, lubricating the rollers reduced the nip angle (Figure 10), maximum pressure (Figure 9), and flake density. Since the values of nip angle and maximum pressure were similar to those of DCPD with critical bulk lubrication, a plausible explanation is the higher sensitivity of wall friction to boundary lubrication. That is, bulk lubrication again provides an internal source of lubricant.
As characterized by the parameter Nc, the sizing rate of flakes appears to be simply related to fracture energy—which is a reasonably expected relationship. For MCC, there seems to be a close correlation between flake density and fracture energy (see Table 2), which is also consistent with expected behavior. There is insufficient data to describe the secondary effects of MgSt, which could either reduce strength by acting as a weak layer between particles or act as a binder. It is almost certain that the low strength of DCPD flakes is due to the elastic deformation accumulated in the release zone, which prevents the formation of bonds between particles. Since organic polymers exhibit elasto-plastic deformation, such deformation is much smaller for MCC. Compared with flakes formed from MCC, the smaller elastic recovery and nip angle of DCPD flakes would be a factor contributing to the lower density of DCPD flakes.
The strength of DCPD tablets was much lower than that of MCC tablets (see Table 3 and Figure 13), which is similar to the flake strength data shown by fracture energy in Table 2. The relative value of accumulated elastic strain may again be the main controlling factor. The insensitivity of DCPD tablet strength to lubrication (see Table 3, Figure 12, and 13a) may be due to the fragmentation properties of DCPD powder particles, which expose fresh surfaces and thus inhibit the potential binding effect of MgSt. Bulk lubrication led to a reduction in the strength of MCC tablets (see Table 3, Figure 12, and 13b), indicating that MgSt acts as a weak boundary layer between particles (Zuurman et al., 1999), thereby reducing the bonding strength (Table 4). Therefore, the effect of lubrication on DCPD tablets and MCC tablets is consistent with the trends observed for flakes, as shown in Table 2.
Future work will focus on comparative studies of binary mixtures to reflect more practical formulations. However, from a practical application perspective, some general trends in the current work may be worth considering. First, for MCC (a major formulation component), under given compaction conditions, the addition of lubricant will reduce the nip angle and maximum roller pressure, which leads to decreased flake strength and improved sizing efficiency, but reduced particle size. Second, the strength of tablets can provide a useful indicator for ranking flake strength and also reflects the efficiency of flake sizing. For example, this method can be applied to study the effect of lubricants.
Studying powder properties, compaction processes, and the properties of flakes, sized granules, and tablets helps to understand the effect of boundary lubricants on formulations. Overall, a coherent yet qualitative interpretation of the data is possible, providing some mechanistic insights and practical implications. Finally, the application of bulk lubrication prevents adhesion during compaction but causes a reduction in the strength and solid fraction of MCC tablets. However, these negative effects were not observed in DCPD. Wall lubrication of both powders minimized the reduction in the tensile strength of the final product.
Silence. Translated. Pharmaceutical Affairs Review. 2022-04-14 06:00
