- Email: [email protected]
Effect of particle size and density on the die fill of powders
European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
Contents lists available at SciVerse ScienceDirect
European Journal of Pharmaceutics and Biopharmaceutics journal homepage: www.elsevier.com/locate/ejpb
Research paper
Effect of particle size and density on the die fill of powders L.A. Mills, I.C. Sinka ⇑ Department of Engineering, University of Leicester, Leicester, UK
a r t i c l e
i n f o
Article history: Received 18 September 2012 Accepted in revised form 22 January 2013 Available online 9 February 2013 Keywords: Powder flow Die fill Microcrystalline cellulose Particle size Density
a b s t r a c t The flow behaviour of powders during the process of die fill was examined. Gravity and suction fill experiments were carried out using a model shoe–die system. Five grades of microcrystalline cellulose were studied to identify the effect of particle size and density on flow. Flowability was quantified using the concept of critical velocity. Under gravity fill, the critical velocity was one order of magnitude higher for powders with large particle size compared to smaller particles. Under suction fill conditions, the critical velocity increased significantly compared to gravity fill, showed no consistent relationship with particle size, and the powders performed more similar to one another. Using high speed video, the gravity and suction fill mechanisms were discussed in the context of air flow and pressure build-up. The effect of shoe velocity, suction velocity and height of the powder in the shoe was explored in more detail. It was shown that one can identify individual contributions from material properties and process parameters to the flow behaviour during die fill; however, the flow performance depends on the inter-relationships between powder characteristics and process parameters. The die fill mechanisms described can be used to assist the optimisation of powder formulation and process design. Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction The manufacturing process of pharmaceutical tablets involves filling the powder formulation into a die followed by compaction using rigid punches. The weight and content uniformity are determined be the consistency of the die fill step. The production rate of a typical pharmaceutical tablet press is of the order of ½ million tablets per hour. The importance of die fill process is evident since it can limit not only product quality but also the productivity of the tabletting process. Initial research into die fill was motivated by other powder pressing industries, such as powder metallurgy, hard metals and ceramics. These industries typically employ single station presses where die fill can be described as follows. The powder is placed into a feed hopper connected to a box-like shoe which translates over the opening of the die. The powder is deposited into the die under the effect of gravity. The process of gravity fill was studied experimentally using a model shoe–die system [1]. One of the unique features of the powder flow in a die fill situation is that the material is deposited in a closed cavity, where the air pressure is increased as more and more material enters the die. High speed video observations [2] showed a complex interplay between powder and air. The shoe–die system was also used for fundamental studies of the flow behaviour of pharmaceutical powders [3] with ⇑ Corresponding author. Department of Engineering, University of Leicester, University Road, Leicester LE1 7RH, UK. Tel.: +44 0116 252 2555. E-mail address: [email protected] (I.C. Sinka). 0939-6411/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ejpb.2013.01.012
experiments conducted in air and vacuum. An extended experimental study [4] using a range of die and shoe geometries together with dimensional analysis led to the development of a flow model for pharmaceutical powders. The results were scaled to rotary tablet press conditions, and it was found that the model under-predicted the flow performance of real powders on real presses. A more detailed examination of the die fill process on a rotary tablet press was undertaken, and the ‘suction fill’ effect was identified [5]. In suction fill, while the feeder is above the opening of the die, the bottom punch is moved down by a cam mechanism creating a suction effect when the powder enters the die. Detailed studies of suction fill were carried out, and die fill was discussed in the context of processing parameters used over the full range of processing steps involved in tablet compaction [6]. Recent experimental research by Mendez et al. [7] reproduced representative conditions of the feed frames of rotary tablet presses and examined the changes of powder properties which are induced by the stirring effect of feed frame. Wu and Guo [8] employed discrete element modelling coupled with computational fluid dynamics to explore complex effects taking place during die fill under gravity and suction conditions. The results of the literature referenced above can be summarised as follows. The key properties of the powders that affect die fill include descriptors of particle size, shape, density and surface properties. These influence the cohesiveness and flowability of the powder mass. The model shoe–die system was used to generate rank orders of powders in terms of performance during die fill; this may or may not be consistent with rankings obtained using standard
643
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
techniques, such as angle of repose, critical orifice diameter, flow rates through funnels and bulk/tapped density measurements. For example, powders consisting of large, round and dense particles flow well in general; however, some pharmaceutical materials with small particle size can also flow surprisingly well under certain conditions. A volume of powder material includes the particles as well as a significant volume of air. Density can therefore be defined in a number of ways. Bulk density (also termed as loose bulk density or apparent density) represents the mass of the powder divided to the total volume (including porosity); this is a property of the powder mass. Particle density (also termed as material density, grain density or full density) is the mass of a particle divided to the volume of the particle and is a property that refers to a particle. Unless otherwise stated, the concept of particle density is used in this paper. Pharmaceutical powders are typically light and cohesive in comparison with other classes of powder materials. Key die fill mechanisms such as gravity and suction fill were identified; however, systematic die fill studies with pharmaceutical specific powders are sparse in the literature. For example, the effect of suction [5] was shown to increase the flowability of a powder by a factor of two; however, this study was limited to observations made using a single excipient and a testing rig where the suction punch was driven by gravity. The purpose of the current paper is to focus on the fundamental aspects of die fill and explore the effect of particle size and particle density systematically for a range of controlled experimental conditions. A common pharmaceutical excipient, microcrystalline cellulose, available in different particle size and density grades required for granulation or direct compaction mixtures is used. The same base material makes it possible to separate size and density effects from complex interactions between material and process parameters. The results present insights into die fill phenomena and can be used in two ways: to guide pharmaceutical formulation design and process development and to facilitate the development and calibration of powder flow models.
2. Materials and methods The powder used in this study is microcrystalline cellulose, manufactured by FMC Biopolymer. Five grades were selected, see Table 1. Four of these powders, Avicel PH-105, PH-101, PH-102 and PH-200, have the same bulk density but differ in terms of particle size. PH102 and PH302 have the same nominal particle size but different bulk density. Table 1 is compiled using the manufacturer’s data. The model shoe–die facility used to perform the die fill experiments is described as follows. The system is shown in Fig. 1 and consists of a fixed die mounted on a die table. A shoe is translated over the die using a servo-pneumatic system controlled from a computer. A suction punch is mounted in the die and is operated by another servo-pneumatic actuator which allows the study of gravity and suction fill mechanisms. The servo-pneumatic
Shoe
Die
Suction punch
Fig. 1. Model shoe die system. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
actuators can be controlled to apply velocities up to 1 m/s and accelerations up to 100 m/s2. The shoe and the die are made of transparent material to enable direct observations of the die fill mechanisms using a high speed video camera (manufactured by Olympus). The experimental system is flexible and can be set-up for a range of die and shoe geometry and testing parameters. In this work, the following geometries were used: Shoe: rectangular, length 65 mm, width 30 mm and height 35 mm. Die: square cross-section, 14 14 mm, height 20 mm. The following testing parameters are used: Shoe velocity: in the range of 25–400 mm/s. To reach the prescribed velocity, the acceleration is set to 50 m/s2. Suction punch velocity: 50, 100 and 400 mm/s. Powder height in shoe: 10, 15 and 20 mm. In a typical experiment, unless otherwise stated, the height of the powder in the shoe is 15 mm, and the suction velocity is 400 mm/s. Each of the powders (Table 1) was tested under the above conditions. Three repeat experiments were carried out to observe the experimental scatter. Powder flow is affected by the initial packing of the particles; therefore, the shoe must be filled in a consistent way for repeatable experiments. In order to help create identical initial conditions, the shoe is shaken five times before it is translated over the die at the prescribed velocity. Each shake involves translating the shoe back and forth for ±5 mm at the maximum acceleration/deceleration of the system. 3. Results and discussion It is instructive to begin with qualitative observations of the flow of the powders based on high speed video recordings.
Table 1 Powder properties. Material
Reference code
Nominal particle size (lm)
Apparent (bulk) density (g/cc)
Applications
Avicel Avicel Avicel Avicel Avicel
PH-105 PH-101 PH-102 PH-200 PH-302
20 50 100 180 100
0.20–0.30 0.26–0.31 0.28–0.33 0.29–0.36 0.35–0.46
Superior compactibility Wet granulation Direct compression Superior flow High density
PH-105 PH-101 PH-102 PH-200 PH-302
644
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
3.1. Observations on flow behaviour High speed video observations of gravity fill and suction fill mechanisms are discussed in turn. In gravity fill the height, the suction punch is set to a position of 20 mm (the die height) below the die table. The shoe translates over the empty die at a prescribed velocity and deposits powder into the die. Snapshots for all five materials tested are presented in Fig. 2. The pictures show the configuration of the powder in the die and shoe for direct comparison of the materials for shoe velocities of 50 mm/s and 100 mm/s. Based on Fig. 2, the following qualitative observations are made:
Powders composed of small particles appear more cohesive while larger particles appear more free flowing. The small particle size material (PH105) at the start of the 50 mm/s experiment presents fractures in the powder mass in the shoe as the powder enters the die. The larger particle size materials (PH102, 200 and 302) form a heap in the shoe. The initial conditioning of the powder (the application of the five shakes) changes the top surface of the powder from flat to a heap which is created by the acceleration involved in the shakes. Best illustrated for PH302 at the end of the 50 mm/s experiment, this heap is not symmetrical, because the powder is thrown towards the back of the shoe during the final acceleration stage to reach the prescribed translational velocity.
Fig. 2. Gravity fill mechanisms. Each row of photos corresponds to the powder labelled on the left. The first two columns from the left correspond to experiments conducted at a shoe velocity of 50 mm/s and the next two columns for shoe velocities of 100 mm/s. For each of the two shoe velocities the first column represents the moment in time when the shoe completely covers the die opening (labelled ‘start’) while the second column a position close to the end of the process (labelled ‘end’). The movement of the shoe is from left to right.
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
Flow regimes identified for other powders such as nose flow and bulk flow [1] are observed. The heap produced by the body forces during acceleration described above is similar to a ‘nose’ which translates over the die opening (all experiments at start). The powder flowing into the die extends the shape of the nose, best illustrated by PH102 or 302 at the start of the 100 mm/s experiment. Bulk flow occurs after the shoe completely covers the die. Here, material continues to detach from the bulk of the powder and fall into the die. This is best illustrated by PH101, 102 and 302 at the end of the 50 mm/s experiment. Intermittent flow, observed in experiments described elsewhere for a wider range of pharmaceutical formulations [3,4], is also observed for the poorly flowing pharmaceutical powders PH105 and PH101. Unlike free flowing streams, in intermittent flow blocks of powder detach from the shoe and fall into the die. As the shoe velocity increases, the amount of powder deposited in the shoe decreases. This is observed for all powders. Quantitative measurements are presented in Section 3. In suction fill, the top punch is positioned level with the die table and remains in this position until the translating shoe (travelling at the prescribed velocity) completely covers the die. At this point, the suction punch moves down at maximum acceleration to reach the prescribed suction velocity which is then maintained until the die height is created, and the punch is decelerated at maximum deceleration to stop at the final die height of 20 mm. Key observations of suction fill mechanisms are discussed for two powders PH105 and 200 which have the smallest and largest particle size from the range. Snapshots are presented in Fig. 3, which is organised similarly to Fig. 2. However, at the start of the suction, the top punch is still level with the die; therefore, in this case, the snapshots labelled ‘start’ correspond to at a later stage to observe the powder entering the die. The suction speed is 400 mm/s. The following qualitative observations are made from Fig. 3: There is no nose flow mechanism; the entire die opening is exposed to powder before the suction punch travels downward.
645
Compared to gravity fill, suction fill is much more efficient. The images presented in Fig. 3 are for shoe velocities of 150 and 250 mm/s, which are much higher than the velocities used for gravity fill in Fig. 2. Nevertheless, due to the suction effect, nearly full dies are obtained for both small and large particle size powders. Gravity fill typically involves a deposition of the powder in the die (best illustrated in Fig. 2 for PH200 at 100 mm/s at the end of the process), whereby the particles entering the die roll on the side of the existing heap, which can lead to segregation. In contrast, suction fill can be visualised as ‘drawing’ of a ‘plug’ of powder from the shoe; this mechanism is less likely to induce segregation. Intermittent flow for the small particle size powder is observed during both gravity and suction fill. In the following sections quantitative results are presented.
3.2. Repeatability of experiments and critical velocity Experimental characterisation of the behaviour of powder materials can be difficult because of sampling issues, the heterogeneous nature of the material and sensitivity to testing conditions. For example, flow measures, particularly for organic powders, depend on the ambient humidity during testing and humidity history experienced by the powder during storage. Prior to testing, the powders used in this work were stored in their original bags sealed by the manufacturer. The packing density is also important particularly for irregular shaped powders where mechanical interlocking between particles can play a significant role in forming arches that affect flowability. The shaking cycles used in our experimental procedure were designed to create repeatable initial packing of the particles; however, it is known that die fill experiments using pharmaceutical powders present large scatter of data [3,4]. The experimental results are quantified in terms of critical velocity and fill ratio. The concepts were introduced for gravity fill by Wu et al. [1] and are defined as follows. If the shoe velocity (v) is sufficiently low, then the die is filled after the shoe passes over the die opening. If the shoe velocity is increased above a certain
Fig. 3. Suction fill mechanisms for PH105 (smallest particle size) and PH200 (largest particle size).
646
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
1.0
1.0
PH 105 gravity
0.8
Fill ratio
Fill ratio
0.8
PH 302 gravity
0.6 0.4 0.2
0.6 0.4 0.2
0.0
0.0 0
50
100
150
200
250
300
350
400
0
50
100
Shoe velocity, mm/s
150
(a)
300
350
400
1.2
PH 105 suction
1.0
PH 302 suction
1.0
0.8
0.8
Fill ratio
Fill ratio
250
(b)
1.2
0.6
0.6
0.4
0.4
0.2
0.2
0.0
200
Shoe velocity, mm/s
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
(c)
(d)
Fig. 4. Repeatability of experiments for the small particle size powder PH105 and large particle size powder PH302 for gravity and suction conditions.
threshold, called critical velocity (vc), then the die remains incompletely filled. Thus, the critical velocity is the maximum velocity of the shoe for which a given die is completely filled with powder after one pass. When incomplete fill occurs, one can define the fill ratio (d) as the ratio between the weight of powder in the partially filled die and the weight of powder in a full die. Gravity and suction fill experiments were repeated three times, and the results for PH105 and PH302 are compared in terms of d in Fig. 4a and b for gravity and suction fill experiments, respectively. Fig. 4 illustrates that the shoe–die system can be used to obtain repeatable data. The scatter of the data is larger for the small particles. However, certain data points for gravity fill experiments for PH302 still present a large scatter due to occasional variations of the material. The scatter for suction fill appears larger particularly for PH105; this is due to the intermittent nature of the flow. When comparing repeatability between gravity and friction, one can also observe that the suction fill is more efficient than gravity, and vc is much larger for suction fill than for gravity fill. The data in Fig. 4a can be represented by a straight line of unity for v < vc and a curve for v > vc. The shape of the curve can be described using the following form:
d¼
m 1þn c
m
ð1Þ
where n is a material dependent parameter. Thus, the flowability of the powder for a particular shoe–die configuration can be defined by two parameters, n and vc, which depend on the powder material. The scatter of data makes it difficult to establish the critical velocity experimentally because a large number of tests should be performed using small increments of shoe velocity. However, Eq. (1) offers an objective method to determine the critical velocity which is described next. Shoe velocities that give a full die are not used for fitting; however, they are used in selecting the data for fitting as follows: all partially full data points to the left of the highest known velocity that gives a full die are excluded for fitting. In other words, only those partial fill data are used for fitting that are above the highest known velocity that gives a full die. These data are fitted using Eq. (1), and vc (and n) is obtained from the fit. In certain situations, vc from the fit can be slightly smaller than the maximum velocity that still gives a full die; nevertheless, the procedure described above represents a practical method for defining critical velocity and will be used in this paper. Eq. (1) lends itself to describe the experimental data generated for gravity fill for pharmaceutical powders as well as a wide range of other powder materials [1]. However, Fig. 4c and d suggests that Eq. (1) may not have the most suitable form for fitting suction fill data. In gravity fill d drops sharply, once vc is exceeded; while in suction fill, the drop is more gradual. The mechanisms of suction fill are discussed in more detail later in this paper; however, for
647
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
simplicity, we use Eq. (1) to define vc in suction fill the same way as in gravity fill.
Table 2 Critical velocity for gravity and suction fill. Material
Gravity fill
3.3. Effect of particle size and material density during gravity fill Fig. 5 shows the gravity fill behaviour of all powders tested. The critical velocity vc and exponent n are presented in Table 2 for each of the powders, where the critical velocities were rounded to the nearest mm/s. It can be seen that the critical velocity correlates with the particle size of the powder, for example, the larger the particle size the higher the critical velocity. However, there is little difference between the behaviour of PH102 and PH200. Also, vc for PH302 is smaller than for PH102 although the bulk density of PH302 is higher.
PH-105 PH-101 PH-102 PH-200 PH-302
Suction fill
n
vc (mm/s)
n
vc (mm/s)
0.3877 0.6000 0.7238 0.7612 0.8873
1 20 50 50 43
2.075 6.254 2.267 1.353 1.230
156 203 196 108 110
1.0
PH105 PH101 PH102 PH200 PH302
The suction fill behaviour for all powders is presented in Fig. 6. The critical velocity and exponent n are listed in Table 2. The following observations are made:
Fill ratio
0.8
3.4. Effect of particle size and material density during suction fill
0.6 0.4 0.2
The critical velocity in suction fill is significantly higher that gravity fill. This is consistent with other results for pharmaceutical powders [5] For the smallest particle size (PH105), the critical velocity during gravity fill was close to zero, while with suction fill, the die fill performance is of the same order as the large particle size powders which are more free flowing. Larger particles (PH200) are expected to be free flowing. In suction fill, however, PH200 is outperformed by all smaller particle powders (PH105, 101, 102). This is because suction fill involves more complex interactions between powder and air, and this affects the observed performance. The effect of air in suction fill is discussed in Section 3.7 after the presentation of the entire set of experimental results. Table 2 shows that powder with higher apparent density (PH302) is significantly outperformed by the lower density PH102. This is evident for both gravity and suction fill experiments. However, at this stage, a general conclusion that higher apparent density reduces flowability should not yet be drawn because the manufacturing process of the higher density PH302 may have altered the structure, morphology and surface properties of the granules, which may be responsible for the observed behaviour. 3.5. Effect of powder height in shoe The initial height of the powder in the shoe is an important process parameter for die fill. The height of powder exerts pressure on the powder at the die opening which acts as a driving force for die 1.0
PH105 PH101 PH102 PH200 PH302
Fill ratio
0.8 0.6 0.4 0.2 0.0
0
100
200
300
400
500
600
Shoe velocity, mm/s Fig. 5. Critical velocity for gravity fill.
700
800
0.0
0
100
200
300
400
500
600
700
800
Shoe velocity, mm/s Fig. 6. Critical velocity for suction fill.
fill; however, the pressure exerted by the weight of the powder can also induce local densification and particle interlocking, which can hinder flow. The escape of air from the die is also affected as discussed in Section 3.7. Studies were conducted for each powder in gravity and suction fill for three initial heights: 10, 15 and 20 mm. The results are plotted in Fig. 7, and the critical velocities are summarised in Table 3. The following observations are made for gravity fill. For gravity fill, the smallest height (10 mm) is most efficient in almost all cases. It is remarkable that for the cohesive PH105 vc = 33 mm/s for 10 mm height while if the powder height in the shoe was increased, then practically, no powder enters the die. The results for all powders can be explained by examining the high speed video data: Small height facilitates nose flow and the tip of the nose is in effect falling into the die during the process. Small height is also advantageous because if all the powder at the tip of the nose enters the die without covering it, air can escape from the die more readily; thus, d is increased. When the die fill mechanisms changes from nose flow to bulk flow during gravity fill, a taller powder bed above the die is less permeable to air escaping from the die, which reduces d. For PH101 and 102, the 15 mm height appears to improve the process most. This can be explained in two ways: Larger height: reduces the efficient nose flow effect. Small height: the material is drawn into the die faster than it is supplied; thus, suction fill looses efficiency. The following observations are made for suction fill. For PH 200 and PH302, the die fill performance is insensitive to the height of the powder in the shoe. It should be noted, however, that the suction fill data in Section 3.2 illustrated that these two powders were outperformed by the smaller particle powders. It can be concluded that suction fill improves the die fill efficiency for large particle
648
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
PH 105 gravity 10mm 15mm 20mm
1.0
0.8 Fill ratio
Fill ratio
0.8 0.6 0.4
0.6 0.4 0.2
0.2 0.0
PH 105 suction 10mm 15mm 20mm
1.0
0
50
100
150
200
250
300
350
0.0
400
0
50
100
150
Shoe velocity, mm/s
PH 101 gravity 10mm 15mm 20mm
1.0
0.6 0.4
0
50
100 150 200 250 Shoe velocity, mm/s
300
350
400
0.6 0.4
0.0 0
400
PH 102 gravity 10mm 15mm 20mm
0.6 0.4 0.2
50
100
150 200 250 Shoe velocity, mm/s
300
350
400
PH 102 suction 10mm 15mm 20mm
1.0 0.8
Fill ratio
0.8
Fill ratio
350
0.2
1.0
0.0
300
PH 101 suction 10mm 15mm 20mm
0.8
0.2 0.0
250
1.0
Fill ratio
Fill ratio
0.8
200
Shoe velocity, mm/s
0.6 0.4 0.2
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
Fig. 7. Effect of powder height in shoe on gravity and suction fill behaviour for all materials. The left column is for gravity fill and the right column is for suction fill. The rows correspond to PH105, PH101, PH102, PH200 and PH302, respectively. Each trendline can be identified using the value of the critical velocities listed in Table 3.
powders by a lesser amount than for smaller particle powders; this is related to air effects which are discussed in Section 3.7. Figs. 6 and 7 include data points with fill ratios larger than unity. This is because the weight of powder in a completely filled die is taken from the gravity fill experiment. The data show that the bulk density of powder in a die filled using suction fill can be larger than the density in the same die filled under the effect of gravity only. The results illustrate that during die fill, there are complex and competing mechanisms at work, and it is not appropriate to draw generalisations made from individual observations. To explore the mechanisms in more detail, suction fill experiment was conducted using a range of suction speeds next.
3.6. Effect of suction velocity Suction fill experiments were performed for PH105, PH102 and PH200. Three suction speeds were employed: 50, 100 and 400 mm/s. The results are presented in Fig. 8 and summarised in Table 4. The following observations are made. For PH105, the critical velocity increases significantly when the suction velocity is increased from 50 mm/s to 100 mm/s. However, there appears to be an optimal suction velocity, as vc decreases when vs is increased further to 400 mm/s. One possible explanation of the observation is that the high vacuum created by high vs early in the process was not fully ‘utilised’ by the poor flowing powder, and the suction effect was
649
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
PH 200 gravity 10mm 15mm 20mm
1.0
0.8
Fill ratio
Fill ratio
0.8 0.6 0.4
0.6 0.4
0.2 0.0
PH 200 suction 10mm 15mm 20mm
1.0
0.2
0
50
100
150
200
250
300
350
0.0
400
0
50
100
PH 302 gravity 10mm 15mm 20mm
1.0
0.6 0.4 0.2 0.0
200
250
300
350
400
PH 302 suction 10mm 15mm 20mm
1.0 0.8
Fill ratio
Fill ratio
0.8
150
Shoe velocity, mm/s
Shoe velocity, mm/s
0.6 0.4 0.2
0
50
100
150
200
250
300
350
400
0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
Shoe velocity, mm/s Fig. 7. (continued)
Table 3 Effect of powder height in shoe on critical velocity for gravity and suction fill. Material
Powder height mm
Gravity fill
Suction fill
n
vc (mm/s)
n
vc (mm/s)
PH-105
10 15 20
0.4904 0.3712 0.1608
33 1 0
1.313 2.923 0.8841
177 152 139
PH-101
10 15 20
0.4580 0.6000 0.5134
31 20 3
0.6092 6.254 1.166
42 203 111
PH-102
10 15 20
0.9843 0.7238 0.7595
43 50 19
1.186 N/A 0.4643
90 N/A 31
PH-200
10 15 20
1.0810 0.7612 0.782
76 50 50
1.292 1.353 1.208
112 108 112
PH-302
10 15 20
0.5624 0.9551 0.7379
52 40 11
0.3937 1.213 1.33
63 111 109
lost rapidly during the process as air from the atmosphere can enter the dies through the clearances; this is discussed in Section 3.7. For PH102 vc increases with vs, while for the large particle size, PH200 high vs does no longer improve vc, perhaps due to high permeability of the powder. The above results suggest that the permeability of the powder may be an important contributing factor to die fill performance, and more research is needed to explore these effects. In suction fill experiments, the weight of the powder in a full die used to calculate d is considered same as in the gravity fill experi-
ment. Data points corresponding to full dies are also indicated in Fig. 8 (PH105) to illustrate that the density of powder in the full die can be influenced by the suction speed. In this case, higher density is achieved at lower suction speeds. 3.7. Remarks on die fill mechanisms and air effects In gravity fill, the air pressure in the die is increased as the powder is introduced. Air flows to the atmosphere through any small
650
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
PH 105 50mm/s 100mm/s 400mm/s
1.0
Fill ratio
0.8 0.6 0.4 0.2 0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s PH 102 50mm/s 100mm/s 400mm/s
1.0
Fill ratio
0.8 0.6 0.4
4. Conclusions
0.2 0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s PH 200 50mm/s 100mm/s 400mm/s
1.0 0.8
Fill ratio
bottom punch as well through the powder in the shoe. This behaviour is illustrated in Fig. 9. As discussed in Section 3.5 when the height of powder in the shoe is increased, d decreases because of air escape. Suction fill is relevant to high speed rotary presses as well as single station presses where the lower punch set can be programmed to move downwards during the die fill process. The stages of suction fill have been described by Motazedian et al. [9] as illustrated in Fig. 10. In Stage 1, the suction punch is level with the top of the die, and the cavity does not yet exist. When the suction punch moves downwards, a low pressure is created in the die (Fig. 10, Stage 2). The atmospheric pressure acting on the top of the powder together with the low pressure created by the punch results in a pressure gradient through the powder bed which facilitates the flow of powder into the die. Air enters the cavity together with/through the powder as well as through the clearances of the system. After the suction punch reaches its final position (Stage 3), the pressure increases in the die and can exceed the atmospheric pressure as more and more powder enters the die. In effect, Stage 3 (Fig. 10) resembles gravity fill.
0.6 0.4 0.2 0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s Fig. 8. Effect of suction speed on die fill for PH105, PH102 and PH200.
Table 4 Effect of suction speed on critical velocity. Suction speed
n
vc (mm/s)
PH105 50 100 400
1.893 2.923 2.059
119 152 134
PH102 50 100 400
2.622 4.546 N/A
97 177 N/A
PH200 50 100 400
2.259 1.397 1.353
100 110 108
gaps and through the powder bed in the shoe. In early stages when nose flow is observed, the die opening is not completely covered, and air can escape easily; this makes nose flow very efficient. Once the die opening is covered, and the bulk flow regime is established the air escapes through the clearances between the shoe, die and
The flow behaviour of powders during die fill is affected by the properties of the powder formulation, and the process parameters applied by machinery. In this paper, the effect of particle size and density was examined in detail for gravity and suction fill using a model shoe–die system and five grades of microcrystalline cellulose from the same manufacturer. A powder with smaller particle size presents more cohesive behaviour and exhibits intermittent flow during die fill, while large particle size promotes free flowing behaviour. For gravity fill, a consistent relationship between particle size and flowability (quantified using the concept of critical velocity) was found, which was expected. However, the effect of particle density was compared using two powders of same nominal size but different bulk density, and it was found that the lower bulk density powder showed better flow behaviour. It should be noted, however, that both are granulated powders using the same primary material; therefore, this effect is likely to originate from the granulation process. The efficiency of nose flow in gravity fill was explained with reference to air escape from the die. Process parameters that prolong the efficient nose flow regime (e.g. height of powder in the shoe) were identified. In a practical situation, there may be other parameters that could be considered, for example, the orientation of dies with complex horizontal cross-section with respect to the shoe. The critical velocity can be increased significantly by employing suction fill. Suction fill involves complex air pressure effects, which for the materials tested appear to overshadow the effect of powder properties such as particle size. Qualitative and quantitative observations coupled with air pressure effect suggest that large particles (which allow air escape more easily than small cohesive particles) perform less efficiently under suction fill conditions and the critical velocities observed are of the same order of magnitude for all powders. The die fill process is influenced by the properties of the powder coupled together with process parameters (such as shoe velocity, suction velocity and powder height in shoe). The results presented above can inform optimisations of both gravity and suction fill. For example, for gravity fill, one can aim to extend the nose flow regime. For suction fill, the suction velocity can be chosen to maximise the beneficial effects of Stage 2 (Fig. 10) and reduce Stage 3. The nominal particle size of the powders tested varied from 20 to 180 lm. This information alone can prompt an expectation that
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
651
Fig. 9. Air flow during gravity fill.
Fig. 10. Air flow during suction fill (adapted from [9]).
that these powders exhibited very different flow behaviours. Indeed, 50-fold differences of the critical velocity were measured under gravity fill. Other flow measures (e.g. based on angle of repose, critical orifice diameter, bulk and tapped density, etc.) can be established for ranking powder flowability. However, the results summarised in Tables 2–4 show that there are die fill conditions under which any one of the powders can outperform all other powders. Notwithstanding the value of all existing powder flow measurement methods, it can be concluded that it is virtually impossible to provide rank classifications for flowability based on powder properties alone. However, in the context of a given process, such as die fill, it is possible to identify the contributing factors and optimise a formulation or process by considering the inter-relationships between material properties and process parameters.
Acknowledgements The microcrystalline cellulose was a gift of FMC Biopolymer, Belgium; sincere thanks to Gregory Thoorens and Bruno Leclercq. This paper was written while the corresponding author was on academic study leave.
References [1] C.-Y. Wu, L. Dihoru, A.C.F. Cocks, The flow of powder into simple and stepped dies, Powder Technol. 134 (1–2) (2003) 24–39. [2] C.-Y. Wu, A.C.F. Cocks, Flow behaviour of powders during die filling, Powder Metall. 47 (20) (2004) 127–136. [3] I.C. Sinka, L.C.R. Schneider, A.C.F. Cocks, Measurement of the flow properties of powders with special reference to die fill, Int. J. Pharm. 280 (1-2) (2004) 27–38.
652
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
[4] L.C.R. Schneider, I.C. Sinka, A.C.F. Cocks, Characterisation of the flow behaviour of pharmaceutical powders using a model die-shoe filling system, Powder Technol. 173 (2007) 59–71. [5] S. Jackson, I.C. Sinka, A.C.F. Cocks, The effect of suction during die fill on a rotary tablet press, Eur. J. Pharm. Biopharm. 65 (2) (2007) 253–256. [6] I.C. Sinka, F. Motazedian, A.C.F. Cocks, K.G. Pitt, The effect of processing parameters on pharmaceutical tablet properties, Powder Technol. 189 (2009) 276–284.
[7] R. Mendez, F.J. Muzzio, C. Velazquez, Powder hydrophobicity and flow properties: effect of feed frame design and operating parameters, AIChE J. 58 (2012) 697–706. [8] C.-Y. Wu, Y. Guo, Numerical modelling of suction fill using DEM/CFD, Chem. Eng. Sci. 73 (2012) 231–238. [9] F. Motazedian, I.C. Sinka, A.C.F. Cocks, The influence of airflow on gravity and suction filling, in: Proceedings of the 2007 International Conference on Powder Metallurgy & Particulate Materials, May 13–16, Denver, Colorado, 2007. ISBN13: 978-0-9793488-5-3 (CD).
Contents lists available at SciVerse ScienceDirect
European Journal of Pharmaceutics and Biopharmaceutics journal homepage: www.elsevier.com/locate/ejpb
Research paper
Effect of particle size and density on the die fill of powders L.A. Mills, I.C. Sinka ⇑ Department of Engineering, University of Leicester, Leicester, UK
a r t i c l e
i n f o
Article history: Received 18 September 2012 Accepted in revised form 22 January 2013 Available online 9 February 2013 Keywords: Powder flow Die fill Microcrystalline cellulose Particle size Density
a b s t r a c t The flow behaviour of powders during the process of die fill was examined. Gravity and suction fill experiments were carried out using a model shoe–die system. Five grades of microcrystalline cellulose were studied to identify the effect of particle size and density on flow. Flowability was quantified using the concept of critical velocity. Under gravity fill, the critical velocity was one order of magnitude higher for powders with large particle size compared to smaller particles. Under suction fill conditions, the critical velocity increased significantly compared to gravity fill, showed no consistent relationship with particle size, and the powders performed more similar to one another. Using high speed video, the gravity and suction fill mechanisms were discussed in the context of air flow and pressure build-up. The effect of shoe velocity, suction velocity and height of the powder in the shoe was explored in more detail. It was shown that one can identify individual contributions from material properties and process parameters to the flow behaviour during die fill; however, the flow performance depends on the inter-relationships between powder characteristics and process parameters. The die fill mechanisms described can be used to assist the optimisation of powder formulation and process design. Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction The manufacturing process of pharmaceutical tablets involves filling the powder formulation into a die followed by compaction using rigid punches. The weight and content uniformity are determined be the consistency of the die fill step. The production rate of a typical pharmaceutical tablet press is of the order of ½ million tablets per hour. The importance of die fill process is evident since it can limit not only product quality but also the productivity of the tabletting process. Initial research into die fill was motivated by other powder pressing industries, such as powder metallurgy, hard metals and ceramics. These industries typically employ single station presses where die fill can be described as follows. The powder is placed into a feed hopper connected to a box-like shoe which translates over the opening of the die. The powder is deposited into the die under the effect of gravity. The process of gravity fill was studied experimentally using a model shoe–die system [1]. One of the unique features of the powder flow in a die fill situation is that the material is deposited in a closed cavity, where the air pressure is increased as more and more material enters the die. High speed video observations [2] showed a complex interplay between powder and air. The shoe–die system was also used for fundamental studies of the flow behaviour of pharmaceutical powders [3] with ⇑ Corresponding author. Department of Engineering, University of Leicester, University Road, Leicester LE1 7RH, UK. Tel.: +44 0116 252 2555. E-mail address: [email protected] (I.C. Sinka). 0939-6411/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ejpb.2013.01.012
experiments conducted in air and vacuum. An extended experimental study [4] using a range of die and shoe geometries together with dimensional analysis led to the development of a flow model for pharmaceutical powders. The results were scaled to rotary tablet press conditions, and it was found that the model under-predicted the flow performance of real powders on real presses. A more detailed examination of the die fill process on a rotary tablet press was undertaken, and the ‘suction fill’ effect was identified [5]. In suction fill, while the feeder is above the opening of the die, the bottom punch is moved down by a cam mechanism creating a suction effect when the powder enters the die. Detailed studies of suction fill were carried out, and die fill was discussed in the context of processing parameters used over the full range of processing steps involved in tablet compaction [6]. Recent experimental research by Mendez et al. [7] reproduced representative conditions of the feed frames of rotary tablet presses and examined the changes of powder properties which are induced by the stirring effect of feed frame. Wu and Guo [8] employed discrete element modelling coupled with computational fluid dynamics to explore complex effects taking place during die fill under gravity and suction conditions. The results of the literature referenced above can be summarised as follows. The key properties of the powders that affect die fill include descriptors of particle size, shape, density and surface properties. These influence the cohesiveness and flowability of the powder mass. The model shoe–die system was used to generate rank orders of powders in terms of performance during die fill; this may or may not be consistent with rankings obtained using standard
643
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
techniques, such as angle of repose, critical orifice diameter, flow rates through funnels and bulk/tapped density measurements. For example, powders consisting of large, round and dense particles flow well in general; however, some pharmaceutical materials with small particle size can also flow surprisingly well under certain conditions. A volume of powder material includes the particles as well as a significant volume of air. Density can therefore be defined in a number of ways. Bulk density (also termed as loose bulk density or apparent density) represents the mass of the powder divided to the total volume (including porosity); this is a property of the powder mass. Particle density (also termed as material density, grain density or full density) is the mass of a particle divided to the volume of the particle and is a property that refers to a particle. Unless otherwise stated, the concept of particle density is used in this paper. Pharmaceutical powders are typically light and cohesive in comparison with other classes of powder materials. Key die fill mechanisms such as gravity and suction fill were identified; however, systematic die fill studies with pharmaceutical specific powders are sparse in the literature. For example, the effect of suction [5] was shown to increase the flowability of a powder by a factor of two; however, this study was limited to observations made using a single excipient and a testing rig where the suction punch was driven by gravity. The purpose of the current paper is to focus on the fundamental aspects of die fill and explore the effect of particle size and particle density systematically for a range of controlled experimental conditions. A common pharmaceutical excipient, microcrystalline cellulose, available in different particle size and density grades required for granulation or direct compaction mixtures is used. The same base material makes it possible to separate size and density effects from complex interactions between material and process parameters. The results present insights into die fill phenomena and can be used in two ways: to guide pharmaceutical formulation design and process development and to facilitate the development and calibration of powder flow models.
2. Materials and methods The powder used in this study is microcrystalline cellulose, manufactured by FMC Biopolymer. Five grades were selected, see Table 1. Four of these powders, Avicel PH-105, PH-101, PH-102 and PH-200, have the same bulk density but differ in terms of particle size. PH102 and PH302 have the same nominal particle size but different bulk density. Table 1 is compiled using the manufacturer’s data. The model shoe–die facility used to perform the die fill experiments is described as follows. The system is shown in Fig. 1 and consists of a fixed die mounted on a die table. A shoe is translated over the die using a servo-pneumatic system controlled from a computer. A suction punch is mounted in the die and is operated by another servo-pneumatic actuator which allows the study of gravity and suction fill mechanisms. The servo-pneumatic
Shoe
Die
Suction punch
Fig. 1. Model shoe die system. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
actuators can be controlled to apply velocities up to 1 m/s and accelerations up to 100 m/s2. The shoe and the die are made of transparent material to enable direct observations of the die fill mechanisms using a high speed video camera (manufactured by Olympus). The experimental system is flexible and can be set-up for a range of die and shoe geometry and testing parameters. In this work, the following geometries were used: Shoe: rectangular, length 65 mm, width 30 mm and height 35 mm. Die: square cross-section, 14 14 mm, height 20 mm. The following testing parameters are used: Shoe velocity: in the range of 25–400 mm/s. To reach the prescribed velocity, the acceleration is set to 50 m/s2. Suction punch velocity: 50, 100 and 400 mm/s. Powder height in shoe: 10, 15 and 20 mm. In a typical experiment, unless otherwise stated, the height of the powder in the shoe is 15 mm, and the suction velocity is 400 mm/s. Each of the powders (Table 1) was tested under the above conditions. Three repeat experiments were carried out to observe the experimental scatter. Powder flow is affected by the initial packing of the particles; therefore, the shoe must be filled in a consistent way for repeatable experiments. In order to help create identical initial conditions, the shoe is shaken five times before it is translated over the die at the prescribed velocity. Each shake involves translating the shoe back and forth for ±5 mm at the maximum acceleration/deceleration of the system. 3. Results and discussion It is instructive to begin with qualitative observations of the flow of the powders based on high speed video recordings.
Table 1 Powder properties. Material
Reference code
Nominal particle size (lm)
Apparent (bulk) density (g/cc)
Applications
Avicel Avicel Avicel Avicel Avicel
PH-105 PH-101 PH-102 PH-200 PH-302
20 50 100 180 100
0.20–0.30 0.26–0.31 0.28–0.33 0.29–0.36 0.35–0.46
Superior compactibility Wet granulation Direct compression Superior flow High density
PH-105 PH-101 PH-102 PH-200 PH-302
644
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
3.1. Observations on flow behaviour High speed video observations of gravity fill and suction fill mechanisms are discussed in turn. In gravity fill the height, the suction punch is set to a position of 20 mm (the die height) below the die table. The shoe translates over the empty die at a prescribed velocity and deposits powder into the die. Snapshots for all five materials tested are presented in Fig. 2. The pictures show the configuration of the powder in the die and shoe for direct comparison of the materials for shoe velocities of 50 mm/s and 100 mm/s. Based on Fig. 2, the following qualitative observations are made:
Powders composed of small particles appear more cohesive while larger particles appear more free flowing. The small particle size material (PH105) at the start of the 50 mm/s experiment presents fractures in the powder mass in the shoe as the powder enters the die. The larger particle size materials (PH102, 200 and 302) form a heap in the shoe. The initial conditioning of the powder (the application of the five shakes) changes the top surface of the powder from flat to a heap which is created by the acceleration involved in the shakes. Best illustrated for PH302 at the end of the 50 mm/s experiment, this heap is not symmetrical, because the powder is thrown towards the back of the shoe during the final acceleration stage to reach the prescribed translational velocity.
Fig. 2. Gravity fill mechanisms. Each row of photos corresponds to the powder labelled on the left. The first two columns from the left correspond to experiments conducted at a shoe velocity of 50 mm/s and the next two columns for shoe velocities of 100 mm/s. For each of the two shoe velocities the first column represents the moment in time when the shoe completely covers the die opening (labelled ‘start’) while the second column a position close to the end of the process (labelled ‘end’). The movement of the shoe is from left to right.
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
Flow regimes identified for other powders such as nose flow and bulk flow [1] are observed. The heap produced by the body forces during acceleration described above is similar to a ‘nose’ which translates over the die opening (all experiments at start). The powder flowing into the die extends the shape of the nose, best illustrated by PH102 or 302 at the start of the 100 mm/s experiment. Bulk flow occurs after the shoe completely covers the die. Here, material continues to detach from the bulk of the powder and fall into the die. This is best illustrated by PH101, 102 and 302 at the end of the 50 mm/s experiment. Intermittent flow, observed in experiments described elsewhere for a wider range of pharmaceutical formulations [3,4], is also observed for the poorly flowing pharmaceutical powders PH105 and PH101. Unlike free flowing streams, in intermittent flow blocks of powder detach from the shoe and fall into the die. As the shoe velocity increases, the amount of powder deposited in the shoe decreases. This is observed for all powders. Quantitative measurements are presented in Section 3. In suction fill, the top punch is positioned level with the die table and remains in this position until the translating shoe (travelling at the prescribed velocity) completely covers the die. At this point, the suction punch moves down at maximum acceleration to reach the prescribed suction velocity which is then maintained until the die height is created, and the punch is decelerated at maximum deceleration to stop at the final die height of 20 mm. Key observations of suction fill mechanisms are discussed for two powders PH105 and 200 which have the smallest and largest particle size from the range. Snapshots are presented in Fig. 3, which is organised similarly to Fig. 2. However, at the start of the suction, the top punch is still level with the die; therefore, in this case, the snapshots labelled ‘start’ correspond to at a later stage to observe the powder entering the die. The suction speed is 400 mm/s. The following qualitative observations are made from Fig. 3: There is no nose flow mechanism; the entire die opening is exposed to powder before the suction punch travels downward.
645
Compared to gravity fill, suction fill is much more efficient. The images presented in Fig. 3 are for shoe velocities of 150 and 250 mm/s, which are much higher than the velocities used for gravity fill in Fig. 2. Nevertheless, due to the suction effect, nearly full dies are obtained for both small and large particle size powders. Gravity fill typically involves a deposition of the powder in the die (best illustrated in Fig. 2 for PH200 at 100 mm/s at the end of the process), whereby the particles entering the die roll on the side of the existing heap, which can lead to segregation. In contrast, suction fill can be visualised as ‘drawing’ of a ‘plug’ of powder from the shoe; this mechanism is less likely to induce segregation. Intermittent flow for the small particle size powder is observed during both gravity and suction fill. In the following sections quantitative results are presented.
3.2. Repeatability of experiments and critical velocity Experimental characterisation of the behaviour of powder materials can be difficult because of sampling issues, the heterogeneous nature of the material and sensitivity to testing conditions. For example, flow measures, particularly for organic powders, depend on the ambient humidity during testing and humidity history experienced by the powder during storage. Prior to testing, the powders used in this work were stored in their original bags sealed by the manufacturer. The packing density is also important particularly for irregular shaped powders where mechanical interlocking between particles can play a significant role in forming arches that affect flowability. The shaking cycles used in our experimental procedure were designed to create repeatable initial packing of the particles; however, it is known that die fill experiments using pharmaceutical powders present large scatter of data [3,4]. The experimental results are quantified in terms of critical velocity and fill ratio. The concepts were introduced for gravity fill by Wu et al. [1] and are defined as follows. If the shoe velocity (v) is sufficiently low, then the die is filled after the shoe passes over the die opening. If the shoe velocity is increased above a certain
Fig. 3. Suction fill mechanisms for PH105 (smallest particle size) and PH200 (largest particle size).
646
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
1.0
1.0
PH 105 gravity
0.8
Fill ratio
Fill ratio
0.8
PH 302 gravity
0.6 0.4 0.2
0.6 0.4 0.2
0.0
0.0 0
50
100
150
200
250
300
350
400
0
50
100
Shoe velocity, mm/s
150
(a)
300
350
400
1.2
PH 105 suction
1.0
PH 302 suction
1.0
0.8
0.8
Fill ratio
Fill ratio
250
(b)
1.2
0.6
0.6
0.4
0.4
0.2
0.2
0.0
200
Shoe velocity, mm/s
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
(c)
(d)
Fig. 4. Repeatability of experiments for the small particle size powder PH105 and large particle size powder PH302 for gravity and suction conditions.
threshold, called critical velocity (vc), then the die remains incompletely filled. Thus, the critical velocity is the maximum velocity of the shoe for which a given die is completely filled with powder after one pass. When incomplete fill occurs, one can define the fill ratio (d) as the ratio between the weight of powder in the partially filled die and the weight of powder in a full die. Gravity and suction fill experiments were repeated three times, and the results for PH105 and PH302 are compared in terms of d in Fig. 4a and b for gravity and suction fill experiments, respectively. Fig. 4 illustrates that the shoe–die system can be used to obtain repeatable data. The scatter of the data is larger for the small particles. However, certain data points for gravity fill experiments for PH302 still present a large scatter due to occasional variations of the material. The scatter for suction fill appears larger particularly for PH105; this is due to the intermittent nature of the flow. When comparing repeatability between gravity and friction, one can also observe that the suction fill is more efficient than gravity, and vc is much larger for suction fill than for gravity fill. The data in Fig. 4a can be represented by a straight line of unity for v < vc and a curve for v > vc. The shape of the curve can be described using the following form:
d¼
m 1þn c
m
ð1Þ
where n is a material dependent parameter. Thus, the flowability of the powder for a particular shoe–die configuration can be defined by two parameters, n and vc, which depend on the powder material. The scatter of data makes it difficult to establish the critical velocity experimentally because a large number of tests should be performed using small increments of shoe velocity. However, Eq. (1) offers an objective method to determine the critical velocity which is described next. Shoe velocities that give a full die are not used for fitting; however, they are used in selecting the data for fitting as follows: all partially full data points to the left of the highest known velocity that gives a full die are excluded for fitting. In other words, only those partial fill data are used for fitting that are above the highest known velocity that gives a full die. These data are fitted using Eq. (1), and vc (and n) is obtained from the fit. In certain situations, vc from the fit can be slightly smaller than the maximum velocity that still gives a full die; nevertheless, the procedure described above represents a practical method for defining critical velocity and will be used in this paper. Eq. (1) lends itself to describe the experimental data generated for gravity fill for pharmaceutical powders as well as a wide range of other powder materials [1]. However, Fig. 4c and d suggests that Eq. (1) may not have the most suitable form for fitting suction fill data. In gravity fill d drops sharply, once vc is exceeded; while in suction fill, the drop is more gradual. The mechanisms of suction fill are discussed in more detail later in this paper; however, for
647
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
simplicity, we use Eq. (1) to define vc in suction fill the same way as in gravity fill.
Table 2 Critical velocity for gravity and suction fill. Material
Gravity fill
3.3. Effect of particle size and material density during gravity fill Fig. 5 shows the gravity fill behaviour of all powders tested. The critical velocity vc and exponent n are presented in Table 2 for each of the powders, where the critical velocities were rounded to the nearest mm/s. It can be seen that the critical velocity correlates with the particle size of the powder, for example, the larger the particle size the higher the critical velocity. However, there is little difference between the behaviour of PH102 and PH200. Also, vc for PH302 is smaller than for PH102 although the bulk density of PH302 is higher.
PH-105 PH-101 PH-102 PH-200 PH-302
Suction fill
n
vc (mm/s)
n
vc (mm/s)
0.3877 0.6000 0.7238 0.7612 0.8873
1 20 50 50 43
2.075 6.254 2.267 1.353 1.230
156 203 196 108 110
1.0
PH105 PH101 PH102 PH200 PH302
The suction fill behaviour for all powders is presented in Fig. 6. The critical velocity and exponent n are listed in Table 2. The following observations are made:
Fill ratio
0.8
3.4. Effect of particle size and material density during suction fill
0.6 0.4 0.2
The critical velocity in suction fill is significantly higher that gravity fill. This is consistent with other results for pharmaceutical powders [5] For the smallest particle size (PH105), the critical velocity during gravity fill was close to zero, while with suction fill, the die fill performance is of the same order as the large particle size powders which are more free flowing. Larger particles (PH200) are expected to be free flowing. In suction fill, however, PH200 is outperformed by all smaller particle powders (PH105, 101, 102). This is because suction fill involves more complex interactions between powder and air, and this affects the observed performance. The effect of air in suction fill is discussed in Section 3.7 after the presentation of the entire set of experimental results. Table 2 shows that powder with higher apparent density (PH302) is significantly outperformed by the lower density PH102. This is evident for both gravity and suction fill experiments. However, at this stage, a general conclusion that higher apparent density reduces flowability should not yet be drawn because the manufacturing process of the higher density PH302 may have altered the structure, morphology and surface properties of the granules, which may be responsible for the observed behaviour. 3.5. Effect of powder height in shoe The initial height of the powder in the shoe is an important process parameter for die fill. The height of powder exerts pressure on the powder at the die opening which acts as a driving force for die 1.0
PH105 PH101 PH102 PH200 PH302
Fill ratio
0.8 0.6 0.4 0.2 0.0
0
100
200
300
400
500
600
Shoe velocity, mm/s Fig. 5. Critical velocity for gravity fill.
700
800
0.0
0
100
200
300
400
500
600
700
800
Shoe velocity, mm/s Fig. 6. Critical velocity for suction fill.
fill; however, the pressure exerted by the weight of the powder can also induce local densification and particle interlocking, which can hinder flow. The escape of air from the die is also affected as discussed in Section 3.7. Studies were conducted for each powder in gravity and suction fill for three initial heights: 10, 15 and 20 mm. The results are plotted in Fig. 7, and the critical velocities are summarised in Table 3. The following observations are made for gravity fill. For gravity fill, the smallest height (10 mm) is most efficient in almost all cases. It is remarkable that for the cohesive PH105 vc = 33 mm/s for 10 mm height while if the powder height in the shoe was increased, then practically, no powder enters the die. The results for all powders can be explained by examining the high speed video data: Small height facilitates nose flow and the tip of the nose is in effect falling into the die during the process. Small height is also advantageous because if all the powder at the tip of the nose enters the die without covering it, air can escape from the die more readily; thus, d is increased. When the die fill mechanisms changes from nose flow to bulk flow during gravity fill, a taller powder bed above the die is less permeable to air escaping from the die, which reduces d. For PH101 and 102, the 15 mm height appears to improve the process most. This can be explained in two ways: Larger height: reduces the efficient nose flow effect. Small height: the material is drawn into the die faster than it is supplied; thus, suction fill looses efficiency. The following observations are made for suction fill. For PH 200 and PH302, the die fill performance is insensitive to the height of the powder in the shoe. It should be noted, however, that the suction fill data in Section 3.2 illustrated that these two powders were outperformed by the smaller particle powders. It can be concluded that suction fill improves the die fill efficiency for large particle
648
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
PH 105 gravity 10mm 15mm 20mm
1.0
0.8 Fill ratio
Fill ratio
0.8 0.6 0.4
0.6 0.4 0.2
0.2 0.0
PH 105 suction 10mm 15mm 20mm
1.0
0
50
100
150
200
250
300
350
0.0
400
0
50
100
150
Shoe velocity, mm/s
PH 101 gravity 10mm 15mm 20mm
1.0
0.6 0.4
0
50
100 150 200 250 Shoe velocity, mm/s
300
350
400
0.6 0.4
0.0 0
400
PH 102 gravity 10mm 15mm 20mm
0.6 0.4 0.2
50
100
150 200 250 Shoe velocity, mm/s
300
350
400
PH 102 suction 10mm 15mm 20mm
1.0 0.8
Fill ratio
0.8
Fill ratio
350
0.2
1.0
0.0
300
PH 101 suction 10mm 15mm 20mm
0.8
0.2 0.0
250
1.0
Fill ratio
Fill ratio
0.8
200
Shoe velocity, mm/s
0.6 0.4 0.2
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
Fig. 7. Effect of powder height in shoe on gravity and suction fill behaviour for all materials. The left column is for gravity fill and the right column is for suction fill. The rows correspond to PH105, PH101, PH102, PH200 and PH302, respectively. Each trendline can be identified using the value of the critical velocities listed in Table 3.
powders by a lesser amount than for smaller particle powders; this is related to air effects which are discussed in Section 3.7. Figs. 6 and 7 include data points with fill ratios larger than unity. This is because the weight of powder in a completely filled die is taken from the gravity fill experiment. The data show that the bulk density of powder in a die filled using suction fill can be larger than the density in the same die filled under the effect of gravity only. The results illustrate that during die fill, there are complex and competing mechanisms at work, and it is not appropriate to draw generalisations made from individual observations. To explore the mechanisms in more detail, suction fill experiment was conducted using a range of suction speeds next.
3.6. Effect of suction velocity Suction fill experiments were performed for PH105, PH102 and PH200. Three suction speeds were employed: 50, 100 and 400 mm/s. The results are presented in Fig. 8 and summarised in Table 4. The following observations are made. For PH105, the critical velocity increases significantly when the suction velocity is increased from 50 mm/s to 100 mm/s. However, there appears to be an optimal suction velocity, as vc decreases when vs is increased further to 400 mm/s. One possible explanation of the observation is that the high vacuum created by high vs early in the process was not fully ‘utilised’ by the poor flowing powder, and the suction effect was
649
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
PH 200 gravity 10mm 15mm 20mm
1.0
0.8
Fill ratio
Fill ratio
0.8 0.6 0.4
0.6 0.4
0.2 0.0
PH 200 suction 10mm 15mm 20mm
1.0
0.2
0
50
100
150
200
250
300
350
0.0
400
0
50
100
PH 302 gravity 10mm 15mm 20mm
1.0
0.6 0.4 0.2 0.0
200
250
300
350
400
PH 302 suction 10mm 15mm 20mm
1.0 0.8
Fill ratio
Fill ratio
0.8
150
Shoe velocity, mm/s
Shoe velocity, mm/s
0.6 0.4 0.2
0
50
100
150
200
250
300
350
400
0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s
Shoe velocity, mm/s Fig. 7. (continued)
Table 3 Effect of powder height in shoe on critical velocity for gravity and suction fill. Material
Powder height mm
Gravity fill
Suction fill
n
vc (mm/s)
n
vc (mm/s)
PH-105
10 15 20
0.4904 0.3712 0.1608
33 1 0
1.313 2.923 0.8841
177 152 139
PH-101
10 15 20
0.4580 0.6000 0.5134
31 20 3
0.6092 6.254 1.166
42 203 111
PH-102
10 15 20
0.9843 0.7238 0.7595
43 50 19
1.186 N/A 0.4643
90 N/A 31
PH-200
10 15 20
1.0810 0.7612 0.782
76 50 50
1.292 1.353 1.208
112 108 112
PH-302
10 15 20
0.5624 0.9551 0.7379
52 40 11
0.3937 1.213 1.33
63 111 109
lost rapidly during the process as air from the atmosphere can enter the dies through the clearances; this is discussed in Section 3.7. For PH102 vc increases with vs, while for the large particle size, PH200 high vs does no longer improve vc, perhaps due to high permeability of the powder. The above results suggest that the permeability of the powder may be an important contributing factor to die fill performance, and more research is needed to explore these effects. In suction fill experiments, the weight of the powder in a full die used to calculate d is considered same as in the gravity fill experi-
ment. Data points corresponding to full dies are also indicated in Fig. 8 (PH105) to illustrate that the density of powder in the full die can be influenced by the suction speed. In this case, higher density is achieved at lower suction speeds. 3.7. Remarks on die fill mechanisms and air effects In gravity fill, the air pressure in the die is increased as the powder is introduced. Air flows to the atmosphere through any small
650
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
PH 105 50mm/s 100mm/s 400mm/s
1.0
Fill ratio
0.8 0.6 0.4 0.2 0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s PH 102 50mm/s 100mm/s 400mm/s
1.0
Fill ratio
0.8 0.6 0.4
4. Conclusions
0.2 0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s PH 200 50mm/s 100mm/s 400mm/s
1.0 0.8
Fill ratio
bottom punch as well through the powder in the shoe. This behaviour is illustrated in Fig. 9. As discussed in Section 3.5 when the height of powder in the shoe is increased, d decreases because of air escape. Suction fill is relevant to high speed rotary presses as well as single station presses where the lower punch set can be programmed to move downwards during the die fill process. The stages of suction fill have been described by Motazedian et al. [9] as illustrated in Fig. 10. In Stage 1, the suction punch is level with the top of the die, and the cavity does not yet exist. When the suction punch moves downwards, a low pressure is created in the die (Fig. 10, Stage 2). The atmospheric pressure acting on the top of the powder together with the low pressure created by the punch results in a pressure gradient through the powder bed which facilitates the flow of powder into the die. Air enters the cavity together with/through the powder as well as through the clearances of the system. After the suction punch reaches its final position (Stage 3), the pressure increases in the die and can exceed the atmospheric pressure as more and more powder enters the die. In effect, Stage 3 (Fig. 10) resembles gravity fill.
0.6 0.4 0.2 0.0
0
50
100
150
200
250
300
350
400
Shoe velocity, mm/s Fig. 8. Effect of suction speed on die fill for PH105, PH102 and PH200.
Table 4 Effect of suction speed on critical velocity. Suction speed
n
vc (mm/s)
PH105 50 100 400
1.893 2.923 2.059
119 152 134
PH102 50 100 400
2.622 4.546 N/A
97 177 N/A
PH200 50 100 400
2.259 1.397 1.353
100 110 108
gaps and through the powder bed in the shoe. In early stages when nose flow is observed, the die opening is not completely covered, and air can escape easily; this makes nose flow very efficient. Once the die opening is covered, and the bulk flow regime is established the air escapes through the clearances between the shoe, die and
The flow behaviour of powders during die fill is affected by the properties of the powder formulation, and the process parameters applied by machinery. In this paper, the effect of particle size and density was examined in detail for gravity and suction fill using a model shoe–die system and five grades of microcrystalline cellulose from the same manufacturer. A powder with smaller particle size presents more cohesive behaviour and exhibits intermittent flow during die fill, while large particle size promotes free flowing behaviour. For gravity fill, a consistent relationship between particle size and flowability (quantified using the concept of critical velocity) was found, which was expected. However, the effect of particle density was compared using two powders of same nominal size but different bulk density, and it was found that the lower bulk density powder showed better flow behaviour. It should be noted, however, that both are granulated powders using the same primary material; therefore, this effect is likely to originate from the granulation process. The efficiency of nose flow in gravity fill was explained with reference to air escape from the die. Process parameters that prolong the efficient nose flow regime (e.g. height of powder in the shoe) were identified. In a practical situation, there may be other parameters that could be considered, for example, the orientation of dies with complex horizontal cross-section with respect to the shoe. The critical velocity can be increased significantly by employing suction fill. Suction fill involves complex air pressure effects, which for the materials tested appear to overshadow the effect of powder properties such as particle size. Qualitative and quantitative observations coupled with air pressure effect suggest that large particles (which allow air escape more easily than small cohesive particles) perform less efficiently under suction fill conditions and the critical velocities observed are of the same order of magnitude for all powders. The die fill process is influenced by the properties of the powder coupled together with process parameters (such as shoe velocity, suction velocity and powder height in shoe). The results presented above can inform optimisations of both gravity and suction fill. For example, for gravity fill, one can aim to extend the nose flow regime. For suction fill, the suction velocity can be chosen to maximise the beneficial effects of Stage 2 (Fig. 10) and reduce Stage 3. The nominal particle size of the powders tested varied from 20 to 180 lm. This information alone can prompt an expectation that
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
651
Fig. 9. Air flow during gravity fill.
Fig. 10. Air flow during suction fill (adapted from [9]).
that these powders exhibited very different flow behaviours. Indeed, 50-fold differences of the critical velocity were measured under gravity fill. Other flow measures (e.g. based on angle of repose, critical orifice diameter, bulk and tapped density, etc.) can be established for ranking powder flowability. However, the results summarised in Tables 2–4 show that there are die fill conditions under which any one of the powders can outperform all other powders. Notwithstanding the value of all existing powder flow measurement methods, it can be concluded that it is virtually impossible to provide rank classifications for flowability based on powder properties alone. However, in the context of a given process, such as die fill, it is possible to identify the contributing factors and optimise a formulation or process by considering the inter-relationships between material properties and process parameters.
Acknowledgements The microcrystalline cellulose was a gift of FMC Biopolymer, Belgium; sincere thanks to Gregory Thoorens and Bruno Leclercq. This paper was written while the corresponding author was on academic study leave.
References [1] C.-Y. Wu, L. Dihoru, A.C.F. Cocks, The flow of powder into simple and stepped dies, Powder Technol. 134 (1–2) (2003) 24–39. [2] C.-Y. Wu, A.C.F. Cocks, Flow behaviour of powders during die filling, Powder Metall. 47 (20) (2004) 127–136. [3] I.C. Sinka, L.C.R. Schneider, A.C.F. Cocks, Measurement of the flow properties of powders with special reference to die fill, Int. J. Pharm. 280 (1-2) (2004) 27–38.
652
L.A. Mills, I.C. Sinka / European Journal of Pharmaceutics and Biopharmaceutics 84 (2013) 642–652
[4] L.C.R. Schneider, I.C. Sinka, A.C.F. Cocks, Characterisation of the flow behaviour of pharmaceutical powders using a model die-shoe filling system, Powder Technol. 173 (2007) 59–71. [5] S. Jackson, I.C. Sinka, A.C.F. Cocks, The effect of suction during die fill on a rotary tablet press, Eur. J. Pharm. Biopharm. 65 (2) (2007) 253–256. [6] I.C. Sinka, F. Motazedian, A.C.F. Cocks, K.G. Pitt, The effect of processing parameters on pharmaceutical tablet properties, Powder Technol. 189 (2009) 276–284.
[7] R. Mendez, F.J. Muzzio, C. Velazquez, Powder hydrophobicity and flow properties: effect of feed frame design and operating parameters, AIChE J. 58 (2012) 697–706. [8] C.-Y. Wu, Y. Guo, Numerical modelling of suction fill using DEM/CFD, Chem. Eng. Sci. 73 (2012) 231–238. [9] F. Motazedian, I.C. Sinka, A.C.F. Cocks, The influence of airflow on gravity and suction filling, in: Proceedings of the 2007 International Conference on Powder Metallurgy & Particulate Materials, May 13–16, Denver, Colorado, 2007. ISBN13: 978-0-9793488-5-3 (CD).