THE EFFECT OF DRY GRANULATION ON FLOW BEHAVIOUR OF

PHARMACEUTICAL POWDERS DURING DIE FILLINGSerena Schiano 1, Lan Chen 2, Chuan-Yu Wu 1

1 Deprtment of Chemical & Process Engineering, University of Surrey, Guildford, GU2 7XH, UK; c.y.wu@surrey.ac.uk

2 School of Medical Instrument and Food Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China;

Abstract

Flowability that quantifies the flow behaviour of powders is an important material attribute

for such applications as packing, hopper flow and powder transport. It is also one of the

critical material attributes of pharmaceutical formulations for solid dosage forms. It is

anticipated that size enlargement via dry/wet granulation will improve the flowability of feed

powders, but it is still unclear how significant the flowability can be enhanced. Therefore, in

this study, an experimental investigation was performed to explore how dry granulation

affects the flowability of pharmaceutical powders, such as microcrystalline cellulose (MCCs),

mannitol and lactose. Both as-received powders and binary mixtures were considered.

Granules of various sizes were produced using roll compaction followed by ribbon milling,

and the flowabiltiy of as-received powders and produced granules was characterised using

two methods: 1) the critical filling speed measured using a model die filling system and 2) the

flow index measured using a Flodex tester. It was shown that the followability increases as

the size of the granules increases for all materials considered. Furthermore, it was found that

there is a strong correlation between the critical filling speed and the flow index: the critical

filling speed is proportional to the flow index to a power of -5/2.

Keywords: Flowability; dry granulation, roll compaction; critical filling speed; die filling

Corresponding Author. Email: C.Y.Wu@surrey.ac.uk. Tel. 0044-1483683506

1

1

2

3

45678

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

1

1. Introduction

In the pharmaceutical industry, it is well recognized that composition variation and the

quality of tablets are determined by material properties and process conditions. One of the

greatest challenges in pharmaceutical development is to identify i) the causal relationship

between material properties, process variables and final product properties, and ii) the critical

material attributes dominating the product properties, which is of practical importance to

obtain high quality products. Pharmaceutical tablets are generally manufactured by

compressing dry powders or granules in a die, i.e., the die compaction, which is the so-called

tabletting process. The tabletting process consists of three primary stages: die filling,

compaction and ejection (Wu et al., 2003). Die filling is a process in which powders are

deposited into a die under the effect of gravity or suction. It is a critical process step during

tabletting, as the flow behaviour during die filling determines tablet properties (e.g. weight

variation, content uniformity), and dictates the segregation tendency of powder blends during

the tabletting process (Wu et al., 2003, Schneider et al., 2007, Wu, 2008).

During die filling, flow behaviour of powders depends on particle density (i.e. true density)

and bulk density. For example, Xie and Puri (2012) investigated the die filling process using

a pressure deposition tester (PDT-II) for alumina powders of different bulk densities. The

tester generated real time pressure distributions of powders at the base of the die during die

filling. It was shown that materials of low bulk density led to irregular and low reproducible

pressure profiles, implying that lower bulk density powders had a higher tendency to produce

non-uniform packing during die filling that led to non-homogeneous tablet density during

tabletting. The influence of particle density during die filling was also examined by Guo et al.

(2011), who explored segregation behaviour of a binary mixture consisting of particles

having the same size but different densities using a coupled discrete element method with

2

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

computational fluid dynamic (DEM-CFD). It was shown that the difference in densities

caused segregation during die filling, in particular, light particles tend to settle on the top of

the packed powder bed while dense particles at the bottom. It was also found that this

tendency was enhanced in the presence of air and the segregation tendency was reduced in

absence of air.

Particle size also affects die filling behaviour. Mills and Sinka (2013) explored the effect of

particle size on gravity and suction filling with different grades of microcrystalline cellulose

powders and found that fine particles showed intermittent flow behaviour due to strong

cohesion, while smooth mass flow was observed for large particles. Wu et al. (2010b)

investigated powder flow behaviour during die filling using the positron emission particle

tracking technique (PEPT) that measured the velocities of individual particles. Two grades of

spherical microcrystalline cellulose powders with different particle sizes were examined

using a model shoe system developed previously (Wu et al., 2003). It was shown that coarse

powders resulted in a higher critical filling speed, indicating that they possess a better

flowability and a higher die filling efficiency can be achieved, comparing to the fine powder.

This was attributed to the presence of air in the die, which can significantly inhibit the flow of

small particles. These observations were consistent with the numerical results of Guo et al.

(2009), who modelled the die filling process in vacuum (i.e. absence of air) and in the

presence of air using DEM-CFD. They found that, during die filling in presence of air, the air

inhibited the flow of powders of small particles and of low density, as no significant

difference was observed for die filling with particles of different sizes and densities in

vacuum.

3

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

Many attempts were also made to understand the correlation between powder flowability and

die filling behaviour. For example, Xie and Puri (2012) argued that fine particles had poor

flowability due to the small size and the increased surface area, which led to an increase of

cohesive forces. Consequently, lower die filling efficiency and poorer content uniformity

were obtained. However, other studies showed that there was not a strong correlation

between powder flowability and filling behaviour. For example, Felton et al. (2002)

investigated capsule filling with mixtures composed of MCC and silicified microcrystalline

cellulose (SMCC) using a tamping-tape encapsulation machine, aiming to understand the

influence of powder flowability on the encapsulation filling. They showed that the fill weight

was higher and more reproducible with SMCC that has better flowability. However, similar

results were obtained for the MCC powder even though their flowability was poorer. They

hence suggested that powder flowability might not be a critical parameter for encapsulation

filling. Wu et al. (2012) evaluated the correlation between flow behaviours of powders

during die filling and flow properties characterised using various methods, include the shear

cell tester, flowmeter, angle of repose, Haunsor index and Carr index, and found that the

flowability testing methods mimicking the powder flow in actual applications gave the better

indication of the ability of powder to flow in the specific application than other powder

flowability tests.

Although die filling has attracted increasing attention in the past two decades, previous

studies have primarily focused on feed powders and the influence of the granule properties on

die filling behaviour has not been investigated. Hence, in this study, the flow behaviour of

granules produced using dry granulation with roll compaction was investigated. The

influence of granule size was examined and the correlation between flow behaviour of

granules and flowability measurements using Flodex was explored.

4

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

2. Materials and Methods

Three commonly used pharmaceutical excipients, microcrystalline cellulose (MCC) of three

different grades: Avicel PH 101, Avicel PH 102 and DG (FMC, Biopolymer, USA), lactose

monohydrate (Granulac 140, Meggle GmbH, Germany) and mannitol (Pearlitol 200 SD,

Roquette, UK) were considered. All the powders are of pure component apart from MCC DG

that is a formulated microcrystalline-based excipient composed of 75% of MCC and 25% of

anhydrous calcium phosphate (FMC BioPolymer, 2014). Three binary mixtures (see Table 1)

of MCC Avicel PH 102 and lactose were also considered. These samples were produced by

mixing the powders in a mixer (TURBULA T2F, Wab, UK) for 15 minutes at a constant

speed of 34 min-1 and named as mixture 1, 2 and 3 based on their compositions.

Table 1. Mixtures composition.

Mixtur

eLactose (%) MCC Avicel PH102 (%)

1 25 75

2 50 50

3 75 25

The same custom-made gravity fed roll compactor as reported in Wu et al. (2010a) was used

to produce ribbons with a roll gap of 1.2 mm at a roll speed of 1 rpm. The ribbons were then

milled into granules using a cutting mill (SM 100, Retsch, Germany) equipped with a 4 mm

mesh size screen at a constant speed of 1,500 rpm. The produced granules were then sieved

5

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

into different granule size cuts (1-90, 90-250, 250-500, 500-1,000, 1,000-1,400, 1,400-2,360

μm).

The flowability of the granules, as-received powders and powder mixtures was characterised

using a Flodex tester (Flodex™, Gradco, UK) that assesses the ability of powder to flow

freely through an orifice in a funnel. The diameter of the smallest orifice through which the

sample can pass three times was considered as the flow index. For each test, around 50g of

sample were poured in the funnel, and the lever device was triggered to open the orifice

quickly to initiate the powder flow. From the Flodex tests, the higher the flow index (the

larger the orifice diameter through which the sample can flow), the poorer the powder

flowability. All tests reported here were repeated three times.

Die filling experiments were then performed with the produced granules, the as-received

powders, and powder mixtures, using a model die filling system (see also Wu et al. 2003),

which consists of a shoe driven by a pneumatic driving unit, a positioning controller unit and

a displacement transducer. Shoe speeds in the range of 10 to 400 mm/s were employed. In

each test, the mass deposited in the die was weighted and the fill ratio was calculated by

δ=mx

mT (1)

where mx is the mass deposited into the die at a certain shoe speed, and mT is the mass in a

completely filled die.

From the definition of the fill ratio adopted by Wu and Cock (2004), the critical filling speed

Vc and the index, n, were determined using

6

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

δ=(V c

V s)

n

(2)

where Vc is the critical filling speed, Vs is the shoe speed and n is a parameter of value

0.8~1.6 for most powders (Wu and Cocks, 2004, Schneider et al. 2007).

3. Results

3.1 Flowability of granules made of pure powders

The flow indices () measured using the Flodex tester for granules made of pure powders are

presented in Fig. 1, in which the flow indices of the as-received powders are also

superimposed. The flow index is referred to as the orifice size through which the material can

flow continuously. Therefore, a lower flow index means that a smaller orifice size is needed

for the material to flow, implying a better flowability. For the as-received powders, MCC

powders have much larger flow index values compared to the mannitol powders, implying

that the mannitol powder has the best flowability among four powders considered. Whereas

the MCC powders have a similar flow index value, close examination reveals that the

flowability of MCC PH102 is slightly better than MCC PH101 and MCC DG. It is also

interesting to note that when the granule is sufficiently large (>1,000 m) the flow index does

not vary with the increase in granule size, indicating it becomes insensitive to the granule

size.

For the granules produced, it can be seen that there is an exponential decrease of the flow

index with the increasing granule size for all the materials investigated. This indicates that

coarse granules possess better flowability than fine ones. This is consistent with the results

obtained by Xie and Puri (2012), who suggested that smaller particles have a larger surface

7

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

area which increases the cohesive forces between particles, thus a poor flowability. It is

interesting to notice that, for MCC, the granules generally have lower flow index values than

the as-received powders, indicating that improved flowability is achieved using dry

granulation. This is attributed to the increase in granule sizes. However, the mannitol

granules generally have higher flow index values than the as-received powder, indicating that

the flowability of granules is poorer than the as-received powder ,even though the granules

are generally larger in size than the as-received powders, which is somehow counter-intuitive.

This is believed to be induced by the change in shape from particle to granules, as the

mannitol particles are almost round while the granules are angular, as demonstrated by Perez-

Gandarillas et al. (2016). This is in broad agreement with the observation of Mellmann et al.

(2013), who also found that spherical particles have better flowability than needled-shaped

and rough particles.

Die filling experiments were also performed using the same materials considered in Fig.1.

For each material, the fill ratios at various filling speeds were determined using Eq. (1).

Figure 2 presents the fill ratio as a function of the filling speed for granules of three different

size cuts: small (0-90 μm), medium (250-500 μm) and large (1.400-2.360 μm), as well as for

the as-received powders. It is clear that the fill ratio decreases with the increase of filling

speed for all the materials investigated. From the definition of fill ratio, a higher fill ratio

implies a higher mass deposited into the die. Hence, figure 2 shows that, as the filling speed

increases, less powder is generally deposited into the die. To achieve a given fill ratio, lower

filling speeds are generally required for finer granules. On the other hand, at the same filling

speed, a higher fill ratio indicates a better flowability, as more powder flows into the die. In

order to obtain the fill ratio, die filling with fine granules has to be run at a lower shoe speeds

8

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

than coarse granules. Therefore, it can be seen in Fig. 2 that coarse granules have better

flowability than fine ones.

The critical filling speed (Vc) and the index n were determined through multi-variate fitting

of the fill ratio data as a function of the filling speed, as shown in Fig.2 for all materials and

granule size cuts considered. In Fig. 2, the lines are fitted curves using Eq.(2). The critical

filling speed Vc as a function of the granule size is then plotted in Fig. 3. It can be seen that,

for all materials considered, as the granule size increases, the filling speed generally increases

until it reaches a plateau when the granule size is large (say >1,000 m). It is also noticed that

with such a large granule size (>1,000 m), the critical filling speeds for MCC PH102 and

MCC DG granules are similar and slightly higher than the other two materials considered.

Since a high critical filling speed indicates a better flowability (as the die can be filled with

such a powder at a higher filling speed), Figure 3 shows that the flowability generally

increases as the granule size increases. However, when the granule size is very larger (>1,000

m), the flowability becomes independent of the granule size. The corresponding n values

are plotted against the granule size in Fig. 4, which shows that the n value for large granules

(> 1,000 m) is slightly higher than that for small granules. It is interesting to observe that the

n values generally vary between 0.8 and 1.4, which is in broad agreement with the

experimental data of Wu and Cocks (2004) and Schneider et al. (2007).

9

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

3.2 Flowability of granules made of binary mixtures

Flow index values for the three binary powder mixtures and their granules of different size

cuts are presented in Fig. 5. It is shown that, for the binary powder mixtures, mixture 1 (i.e.

75% MCC and 25% lactose) has the highest flow index value, while mixture 3 has the lowest

one. This implies that for powder mixtures with a higher MCC percentage, a larger orifice

size is needed to achieve a steady powder flow; hence it has a poorer flowability. For all the

granules considered, a similar behaviour to that shown in Fig.1 was observed, i.e. when the

granule is small (< 1,000 m), the flow index decreases with increased granule size for all the

mixtures; when the granular is sufficiently large (>1,000 m), a similar flow index value is

obtained for all three mixtures. This further confirms that increasing granule size generally

improves the flowability, but further increase in the granule size has a limited effect on the

flow index when the granule size is large (say >1,000 m).

Figure 6 shows the fill ratio as a function of the filling speed for granules made of the binary

mixtures in three different size cuts: small (0-90 μm), medium (250-500 μm) and large (1400-

2360 μm). The same features as observed for pure powders (Fig.2) are also observed for

binary mixtures: the fill ratio decreases exponentially with the increase of the filling speed; at

the same filling speed, the fill ratio is higher for larger granules, indicating an improved

flowability with large granules. The critical filling speed (Vc) and the index n for various

sized granules made of three binary mixtures are shown in Figs 7 & 8, respectively. Again, it

is shown that the critical filling speed increases as the size increases, but becomes insensitive

to the granule size when the granules become too large (say >1,000 m). a close examination

of Fig.7 also reveals that a higher critical filling speed is generally obtained for the granules

made of higher fraction of MCC. The index n (see Fig.8) increases as the granule size

10

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

increases and it varies in the range of 0.8 ~1.6, which is in broad agreement with the value

obtained previously.

4. Discussions

4.1 The effect of particle/granule size on flowability

It is recognised that powder flowability can be improved through size enlargement,

because the large particles generally have a smaller specific surface area (i.e. surface area

per unit mass) and lower inter-particle cohesive forces, comparing to the small particles

(Xie and Puri, 2012). It is hence expected that the granules produced from wet or dry

granulation processes will have improved flowability compared to the feed powders.

However, limited evidence was provided in the literature to substantiate this argument,

especially for granulated produced with dry granulation. In the present study, using 4 as-

received powders and 3 binary mixtures, granules were produced with roll compaction

and ribbon milling, and the effect of granule size was evaluated by dividing the granules

into various size cuts. The flowability of these granules was evaluated using two different

techniques: 1) the Flodex tester, which determines the critical orifice size (flow index)

through which the materials can flow; and 2) the critical filling speed, i.e. the highest

speed at which the die can be completely filled in one passage. The Flodex results (Figs 1

& 5) showed that a smaller flow index is obtained with larger granules, confirming that

the flowability was generally improved by granulating the fine powder into large

granules. Similar conclusion can also be drawn from the analysis of the critical filling

speed (Figs 3 & 7), as the critical filling speed increased as the granule size increases.

11

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Nevertheless, two interesting phenomena were also observed in the present study: 1)

when the granule is very large (say >1,000 m), further increase in the granule size offers

limited advantage in improving the flowability, as both the flow index (Figs 1 & 5) and

critical filling speed (Figs 3 & 7) remained unchanged as the granule size increased

further. This is believed to be due to the diminished effect of inter-particle forces when

the granules are sufficient large. 2) Particle shape can also play an important role in

powder flow behaviour, as demonstrated with the mannitol powder. It clearly showed

that, for mannitol, although the granules are generally larger than the as-received feed

powder the feed powder can be discharged through a smaller orifice size than the

granules. This is primarily due to the change in particle/granule shape in dry granulation.

It should also be noted that the effect of particle shape appears dependent on the actual

flow conditions. As shown in Fig.2d, during die filling, the as-received mannitol powders

only had a better flowability than the granules of small size (< 500 m), as a higher

critical filling speed than that of as-receive mannitol powders was obtained when the

granules are large. Hence a more comprehensive understanding of particle shape on

powder flowability deserves further study.

4.2 Correlation between flow index and critical filling speed

As argued by Wu et al. (2003), powder flow behaviour during the die filling process,

which was used to determine the critical filling speed in the present study, is to some

extent similar to hopper flow as happened in the Flodex measurement. For all the

materials used and the orifice sizes considered in the Flodex measurement, the mass flow

rate during the discharge can be approximated using the Beverloo equation (Seville and

Wu, 2016):

12

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

M=C ρb g1 /2 ( D−kd )5/2 (3)

where M is the mass flow rate, C is the Beverloo constant, D is the orifice size, d is the

particle/granule size, k is a parameter related to particle shape and g is the gravitational

acceleration. As the Flodex tester determines the smallest discharge orifice size ψ , the

lowest mass flow rate of the powder during hopper discharge can be given as

M min=C ρb g1 /2 (ψ−kd )5 /2 (4)

The flowability of the powder, Φ, can then be expressed as

Φ=f ( M min )= aM min

= aC ρb g1 /2 (ψ−kd )5 /2 (5)

where a is an empirical parameter. Equation (5) indicates that a smaller value of the flow

index ψ represents a high value of Φ, i.e. better flowability. The flowability of the

powder, Φ, can be interpreted as the shortest time to discharge a powder or pack the

powder into a container/die.

13

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

For the die filling process at the critical filling speed Vc, the effective discharge/filling

time τ can be approximated as

τ= LV c

(6)

where L is the effective length for die filling, which is related to the length of the shoe

and the die. To quantify the flowability of powder Φ in this process, the flowability Φ

can be related to the effective discharge time as

Φ=f ( τ )=bτ=

bV c

L(7)

where b is a constant. Equation (7) indicates that the shorter the effective discharge time

required to fully fill the die, the better the powder flowability is. Moreover, it shows that

the higher the critical filling speed, the better the flowability.

Combining Eqs (5) and (7) leads to

V c=aL

Cb ρb g1 /2 (ψ−kd )5 /2 (8a)

14

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

For d ≪ψ , Eq.(8) can be simplified to

V c=aL

Cb ρb g1 /2 (ψ )5/2 (8b)

Equation (8) indicates that the critical filling speed is proportional to ψ−5 /2, i.e.

V c=α ψ−5 /2 (9)

Using the data reported in Figs 1, 4, 5 and 7, a plot of the critical filling speed against the

flow index ψ is shown in Fig. 9. It can be seen that the data can be well approximated

using Eq.(9), indicating that there is a strong correlation between the critical filling speed

Vc and the flow index ψ. On the other hand, Figure 9 demonstrates that Eq.(9) can well

describe the interdependence of Vc with the flow index ψ.

5. Conclusions

In order to explore whether dry granulation can improve the flowability of feed materials,

flowability of feeds powders, their mixtures, and the granules of various sizes, were evaluated

15

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

using a Flodex tester and a model die filling system, from which the flow index and the

cirtical filling speed were determined, respectively. It was found that coarser granules

generally have a better flowability than fine ones, and the flowability can generally be

improved when the granules are sufficiently large (say > 500 m). For small granules (< 500

m), the effect of particle shape can be dominant as the round feed particles appears to have a

better flowability than the angular granules, even though the latter may have a larger size than

the former. It was also observed that when the granule size is very large (> 1,000 m), the

flowability became insensitive to the granule size. Furthermore, it was found that there is a

strong correlation between the critical filling speed and the flow index. Both experimental

and theoretical analyses showed that the critical filling speed is proportional to the flow index

to a power of -5/2.

ACKNOWLEDGEMENTS

This work was supported by the IPROCOM Marie Curie initial training network, funded

through the People Programme (Marie Curie Actions) of the European Union's Seventh

Framework Programme FP7/2007-2013/ under REA grant agreement No. 316555. The

authors are grateful to FMC Chemicals prl, Brussels, Belgium, for the gift of MCC powders

used in this study.

References

Felton, L.A., Wiley, C.J., and Friar, A.L., 2002. Enteric coating of gelatine and cellulosic capsules using an aqueous-based acrylic polymer. AAPS Pharm. Sci. 4, T3320.

FMC ByoPolymer, 2014. In: Application Bulletin: Avicel® DG Specialty Binder for Dry Granulation.

Guo Y., Kafui K.D., Wu C.Y., Thornton C. and Seville J.P.K., 2009. A coupled DEM/CFD analysis of the effect of air on powder flow during die filling. AICHE Journal, 55 (1): 49-62.

Guo, Y., Wu, C.-Y., Kafui, K.D., and Thornton, C., 2011. 3D DEM/CFD analysis of size-induced segregation during die filling. Powder Tech. 206, 177-188.

16

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

212223242526272829303132

Mellmann, J., Hoffmann, T., and Furll, C., 2013. Flow properties of crushed grains as a function of the particle shape. Powder Tech. 249, 269-273.

Mills, L.A. and Sinka, I.C., 2013. Effect of particle size and density on the die fill of powders. Eur. J. Pharm. Biopharm. 84, 642-652.

Perez-Gandarillas, L., Perez-Gago A., Mazor, A., Kleinebudde, P., Lecoq, O., Michrafy, A., 2016. Effect of roll-compaction and milling conditions on granules and tablet properties, European Journal of Pharmaceutics and Biopharmaceutics, 106, 38-49,

Seville J.P.K. and Wu C.Y., (2016) Particle Technology and Engineering. Elsevier, London.

Schneider, L.C.R., Sinka, I.C., and Cocks, A.C.F., 2007. Characterization of the flow behaviour of pharmaceutical powders using a model die-shoe filling system. Powder Tech. 173, 59-71.

Wu, C.Y., 2008. DEM simulations of die filling during pharmaceutical tabletting. Particuology 6, 412-418.

Wu, C.Y. and Cocks, A.C.F., 2004. Flow behavior of powders during die filling. Powder Metall. 47, 127-136.

Wu, C.Y., Dihoru, L., and Cocks, A.C.F., 2003. The flow of powder into simple and stepped dies. Powder Tech. 134, 24-39.

Wu, C.Y., Armstron, B., Vlachos, N., 2012. Characterisation of Powder flowability for Die filling. Particulate Science and Technology, 30(4): 378-389.

Wu, C.-Y., Hung, W.-L.; Miguélez-Morán, A.M.; Gururajan, B. and Seville, J.P.K., 2010a. Roller compaction of moist pharmaceutical powders. Int. J. Pharm. 391, 90-97.

Wu C-Y, Fan XF, Motazedian F, Seville JPK, Parker DJ, Cocks ACF. 2010b. Quantitative investigation of powder flow during die filling using positron emission particle tracking. PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART E-JOURNAL OF PROCESS MECHANICAL ENGINEERING, 224 (E3), pp. 169-175.

Xie, X., and Puri, V.M., 2012. Powder deposition in three parallel-oriented dies of cylindrical and E shapes. Adv. Powder Technol. 23, 1-7.

17

123456789

10111213141516171819202122232425262728293031323334353637383940

Fig. 1 Flow index () as a function of granule size for granules made of four different

materials.

18

1

2

3

4

(a)

(b)

19

1

2

3

4

(c)

(d)

Fig. 2 Fill ratio as a function of shoe speed for three different granule size cuts made of (a)

MCC PH 102, (b) MCC PH 101, (c) MCC DG and (d) Mannitol.

20

1

2

3

4

5

6

Fig. 3 Critical filling speed as a function of the granule size for various materials considered.

21

1

2

3

4

Fig. 4 The index n as a function of the granule size for various materials considered.

22

1

2

3

Fig. 5 Flow index () as a function of granule size for three binary powders.

23

1

2

3

(a)

(b)

24

1

2

3

4

(c)

Fig. 6 Fill ratio as a function of filling speed for granules of three different size cuts: small

(0-90μm), medium (250-500μm) and large (1400-2360μm), made of (a) Mixture 1, (b)

Mixture 2 and (c) Mixture 3.

25

1

2

3

4

5

6

Fig. 7 Critical filling speed as a function of granule size for granules made of different binary

mixtures.

26

1

2

3

4

Fig. 8 Parameter n as a function of granule size for granules made of different binary

mixtures.

27

1

2

3

4

5

6

Fig. 9 Critical filling speed Vc as a function of Flow index (ψ) for granules made of pure

powders and binary mixtures.

28

1

2

3

4