Linking flowability and granulometry of lactose powders

International Journal of Pharmaceutics 494 (2015) 312–320

Contents lists available at ScienceDirect

International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm

Linking flowability and granulometry of lactose powders F. Boschini a,c , V. Delaval a , K. Traina a,d , N. Vandewalle a,b , G. Lumay a,b,∗ a

APTIS, Institute of Physics, University of Liege, Sart-Tilman, B-4000 Liège, Belgium GRASP, Institute of Physics, University of Liege, B-4000 Liège, Belgium c GreenMat, Institute of Chemistry, University of Liege, B-4000 Liège, Belgium d Galephar MF, R&D Center, B-6900 Marche en Famenne, Belgium b

a r t i c l e

i n f o

Article history: Received 15 April 2015 Received in revised form 9 August 2015 Accepted 10 August 2015 Available online 15 August 2015 Keywords: Lactose Flowability Rheology Packing fraction Cohesion

a b s t r a c t The flowing properties of 10 lactose powders commonly used in pharmaceutical industries have been analyzed with three recently improved measurement methods. The first method is based on the heap shape measurement. This straightforward measurement method provides two physical parameters (angle of repose ˛r and static cohesive index  r ) allowing to make a first screening of the powder properties. The second method allows to estimate the rheological properties of a powder by analyzing the powder flow in a rotating drum. This more advanced method gives a large set of physical parameters (flowing angle ˛f , dynamic cohesive index  f , angle of first avalanche ˛a and powder aeration %ae ) leading to deeper interpretations. The third method is an improvement of the classical bulk and tapped density measurements. In addition to the improvement of the measurement precision, the densification dynamics of the powder bulk submitted to taps is analyzed. The link between the macroscopic physical parameters obtained with these methods and the powder granulometry is analyzed. Moreover, the correlations between the different flowability indexes are discussed. Finally, the link between grain shape and flowability is discussed qualitatively. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Granular materials and fine powders are widely used in pharmaceutical applications (Ashurst et al., 2000; Warnke et al., 1996; Kaialy and Nokhodchi, 2015; Tavares Cardoso et al., 2011) as excipient or active ingredient in formulations. As excipient, lactose powders are involved in many processes: tabletting (Keleb et al., 2004), wet and dry granulation (Braumann et al., 2010; Kumar et al., 2014), blending, caps filling, etc. Therefore, any progress in the understanding of lactose powders flowing behaviors can have huge consequences for pharmaceutical industries. Indeed, a powder with inappropriate flowing properties can cause serious complications in production lines (clogging, agglomeration, segregation, etc.). To control and to optimize processing methods, these materials have to be precisely characterized. The characterization methods are related either to the properties of the grains and to the behavior of the powder bulk. Unfortunately, the relation between the grain properties and the powder behavior is far to be obvious (de Gennes, 1999; Lumay et al., 2012). Therefore, both measurement

∗ Corresponding author at: APTIS, Institute of Physics, University of Liege, SartTilman, B-4000 Liège, Belgium. E-mail address: [email protected] (G. Lumay). http://dx.doi.org/10.1016/j.ijpharm.2015.08.030 0378-5173/© 2015 Elsevier B.V. All rights reserved.

types have to be performed. Many advanced methods are available to measure the grain characteristics: laser diffraction to obtain the grain size distribution (Tinke et al., 2009), granulomorphometer to measure the grain shape, X-ray diffractometer to characterize the grain crystallinity, SEM microscopy to visualize the surface and the shape of the grains, . . . However, concerning the physical behavior of powder bulk (flowability, packing fraction, etc.), most of the techniques used in R&D or quality control laboratories are based on old measurement techniques (European Pharmacopoeia). During the last decade, interesting techniques have been developed like shear cells (Schulze, 2011; Saw et al., 2014) and powder rheometers inspired by liquid rheometers (Freeman, 2007). However, the evolution of this field is still at its beginning. Indeed, even from a fundamental point of view, the determination of the physical laws that govern the behavior of a granular material is still a matter of intense debates in the physics community. In previous studies (Lumay et al., 2012; Traina et al., 2013), we have shown how classical powder characterization techniques (angle of repose, tapped density and rotating drum) can be updated to meet the present requirements of R&D laboratory and production department. In particular, we have shown that the automatization of the measurements and the development of rigorous initialization methods is necessary to obtain reproducible and interpretable results. Moreover, the use of image analysis techniques is a

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

considerable added value for flowability measurement techniques. Indeed, the quality and the quantity of informations extracted from the measurement are improved. In addition, the inspection of the resulting images by the operator provides additional informations. In this paper, we show how three recently improved flowability measurement methods can be used practically to characterize the physical behavior of lactose powders. Moreover, the correlations between the flowability indexes obtained with the different methods are discussed. Finally, the results are correlated with the grain size distributions and grain shapes. 2. Materials and methods

313

Table 2 Definition of the quantities measured with GranuHeap, GranuDrum and GranuPaq instruments. Quantity

Definition

Instrument

˛r r ˛f f ˛a %ae (0) (500) n1/2 Hr (∞)

Repose angle Static cohesive index Flowing angle Dynamic cohesive index First avalanche angle Powder aeration during a flow Initial density Density after 500 taps Characteristic tap number Hausner ratio Extrapolated optimal density

GranuHeap GranuHeap GranuDrum GranuDrum GranuDrum GranuDrum GranuPaq GranuPaq GranuPaq GranuPaq GranuPaq

2.1. Materials The lactose powders analyzed in the present study are produced by the company Meggle and are widely used in pharmaceutical industries. The powders are separated in four groups corresponding to different production methods and different areas of application: Tablettose for tabletting processes, Granulac for granulation processes, Inhalac for dry powder inhaler applications and Flowlac for direct compression processes. The information provided by Meggle are summarized in Table 1. All the sample are ˛-lactose monohydrate powders. The Tablettose powders are obtained from agglomeration. Granulac powders come from milling process. Inhalac powders are sieved and milled lactose powders. Finally, Flowlac powders are obtained from spray-dried lactose suspensions. Flowlac powders are optimized to have a high flowability and compressibility. 2.2. Methods To evaluate the rheological properties of the powder samples, three recently developed experimental set-ups were used (Lumay et al., 2012): (i) GranuHeap to measure static properties (angle of repose ˛r and static cohesive index  r ), (ii) GranuDrum to measure the flowing properties (flowing angle ˛f , dynamic cohesive index  f , first avalanche angle ˛a and powder aeration (%ae ) during a flow) and (iii) GranuPaq to measure quasi-static properties (Hausner ratio Hr, bulk density (0), tapped density (500) and densification characteristic time n1/2 ). Table 2 summarizes the quantities measured with GranuHeap, GranuDrum and GranuPaq instruments. In order to check the robustness of the measurement methods, the measurements have been repeated three times with three selected samples. These samples have been selected after a first measurement with GranuDrum instrument and correspond to (i) the lactose powder with the higher cohesive index, (ii) the lactose powder with the lower cohesive index and (iii) the powder with an intermediate cohesive index.

During the measurements, the relative humidity in the lab was ranging from 30%RH to 40%RH and the temperature was fixed to 20 ◦ C. Any preconditioning process was applied to the powder samples before the measurements. 2.2.1. Heap shape When a powder is poured onto a surface, a heap is formed. It is well known that both the repose angle ˛r and the heap shape strongly depend on grain properties. In particular, the heap shape depends on the powder cohesiveness (de Ryck et al., 2005, 2010). A cohesive powder forms an irregular heap (see Fig. 1c) while a non-cohesive powder forms a regular conical heap (see Fig. 1a). Therefore, a precise measurement of the heap shape provides useful information about the physical properties of the powder sample. Unfortunately, the final heap shape is highly sensitive to its formation method, in particular with cohesive powders. Therefore, an automated initialization protocol and a precise measurement method have to be defined. GranuHeap instrument (Lumay et al., 2012) is an automated heap shape measurement method based on image processing and analysis. A powder heap is created on a cylindrical support. In order to obtain reproducible results, an initialization tube with an internal diameter equal to the circular support is installed on the support. After filling the initialization tube by hand with a fixed volume of powder (100 ml for the present study), the tube goes up at the constant speed of 5 mm/s. Thereby, the powder is flowing from the tube to form a heap on the cylindrical support. A controlled rotation of the support allows obtaining different heap projections corresponding to different heap orientations. In the present study,

Table 1 Technical data provided by the producer Meggle about the powders: bulk density B , tapped density tap , Hausner ratio Hr and mean grain size d(0.5) expressed in ␮m. Powder

Code

B

tap

Hr

d(0.5)

Tablettose 70 Tablettose 80 Tablettose 100

T70 T80 T100

0.53 0.62 0.58

0.64 0.77 0.72

1.21 1.24 1.24

NC NC NC

Granulac 70 Granulac 140

G70 G140

0.71 0.63

0.91 0.89

1.28 1.41

NC NC

Inhalac 70 Inhalac 120 Inhalac 230

I70 I120 I230

0.63 0.73 0.71

0.72 0.83 0.85

1.14 1.14 1.20

215 132.5 90

Flowlac 90 Flowlac 100

F90 F100

0.56 0.59

0.67 0.71

1.20 1.20

NC NC

Fig. 1. Pictures of the heaps (a–c) and of flow inside the rotating drum (d–f) obtained with three lactose powders. The Inhalac 70 (a,d) is a non-cohesive powder while the Granulac 140 (c,e) is a cohesive powder. The cohesiveness of the Inhalac 230 (b,e) powder is intermediate.

314

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

16 images separated by a rotation of 11.25◦ were recorded. A custom image recognition algorithm determines the position of the powder/air interface. The repose angle ˛r refers to the angle of the isosceles triangle with the same surface than the powder heap projected image. This isosceles triangle corresponds to the ideal cohesiveness heap shape. The repose angle ˛r is computed for each image, i.e. for each heap orientation. Afterward, an averaged value is computed. In general, the lower the repose angle is, the is better the powder flowability (Lumay et al., 2012). However, this kind of general principle should be taken carefully. Indeed, the repose angle is a static parameter. During a flow, some complex mechanisms (powder aeration, formation or breaking of agglomerates, etc.) can influence the flowability. Therefore, the link between a static parameter and dynamical behaviors is often complex. The deviation between the real heap interface and the isosceles triangular heap provides the static cohesive index  r . The static cohesive index  r is computed for each image, i.e. for each heap orientation. Afterward, an averaged value is computed. This static cohesive index  r is close to zero for a non-cohesive powder and increases when the cohesive forces inside the powder strengthen. The advantage of the automated heap shape analysis is the measurement fastness and the small number of extracted parameters (the repose angle ˛r and the static cohesive index  r ). However, to obtain more informations about the powder rheology, a dynamic test like the rotating drum is necessary. 2.2.2. Rotating drum Experimentally, the most practical geometry to study the flow of granular materials is the rotating drum. This flow geometry has been extensively studied fundamentally with non-cohesive granular materials (Rajchenbach, 1990; Fischer et al., 2009; Taberlet et al., 2006) and cohesive granular materials (Chaudhuri et al., 2006; Quintanilla et al., 2006; Lumay and Vandewalle, 2010; Pirard et al., 2009). A horizontal cylinder with glass side walls called drum is half filled with the sample of powder. The cylinder rotates around its axis and the powder flow is analyzed by image processing. This instrument evaluates powder flowing properties for different rotating speeds, i.e. for different shear stresses. Comparing to powder rheometers based on shear cell measurement or on the rotation of a blade, the only stress applied on the powder sample in the rotating drum is induced by gravity. Therefore, the rotating drum method evaluates powder flowing properties in the free flow regime. GranuDrum instrument (Lumay et al., 2012) is an automated powder flowability measurement method based on the rotating drum principle. The drum rotates around its axis at an angular velocity ranging from 2 rpm to 20 rpm. A CCD camera takes snapshots (50 images separated by 0.5 s) for each angular velocity. The air/powder interface is detected on each snapshot with an edge detection algorithm. Afterward, the average interface position and the fluctuations around this average position are computed. Then, for each rotating speed, the flowing angle ˛f is computed from the average interface position and the dynamic cohesive index  f is measured from the interface fluctuations. In general, a low value of the flowing angle ˛f corresponds to a good flowability. However, this kind of general principle should be taken carefully. Indeed, the flowing angle is influenced by a wide set of parameters: the friction between the grains, the shape of the grains, the cohesive forces (van der Waals, electrostatic and capillary forces) between the grains. The dynamic cohesive index  f is only related to the cohesive forces between the grains. A cohesive powder leads to an intermitted flow (see Fig. 1f) while a non-cohesive powder leads to a regular flow (see Fig. 1d). Therefore, a dynamic cohesive index close to zero corresponds to a non-cohesive powder. When the powder cohesiveness increases, the cohesive index increases accordingly. Table 3 summarizes the empirical relation between the cohesive index value and the powder cohesiveness. In addition, the rotating

Table 3 Empirical relation between the flow properties and the results obtained with two well-known powder tests (repose angle and Hausner ratio measurements) (European pharmacopoeia, in press). The corresponding values of the dynamic cohesive index  f are also presented. Flow

˛r (◦ )

Hr

f

Excellent Good Fair Passable Poor Very poor Very very poor

25–30 31–35 36–40 41–45 46–55 56–65 >66

1.00–1.11 1.12–1.18 1.19–1.25 1.26–1.34 1.35–1.45 1.46–1.59 >1.60

<5 5–10 10–20 20–30 30–40 40–50 >50

drum method gives the opportunity to study rheological properties of the powders (shear thinning, shear thickening and thixotropic behavior) by varying the rotating speed. In addition to the measurement of both the cohesive index  f and the flowing angle ˛f as a function of the rotating speed, GranuDrum allows to measure the first avalanche angle ˛a and the powder aeration (%ae ) during the flow. To estimate the first avalanche angle ˛a the drum is initially positioned to obtain an horizontal powder/air interface. Afterward, the drum is rotating slowly till the first avalanche. The first avalanche angle ˛a is the angle of the powder/air interface just before the first avalanche. To measure the powder aeration %ae during the flow, the volume of the powder sample is measured before the rotation V0 and after the rotation Vf of the drum. The rotating speed and time can be adjusted. The powder aeration is calculated with the relation %ae = 100 (Vf − V0 )/V0 . 2.2.3. Density The bulk density, the tapped density and the Hausner ratio measurement are very popular for powder characterization because of both the simplicity and the rapidity of the measurement. The recommended procedure is the following (European Pharmacopoeia). A powder sample is gently poured in a 250 ml glass cylinder. Then, the initial volume V0 of the powder is measured by naked eyes. Afterward, the pile experiences 500 calibrated taps and the final tapped volume Vf is measured. The Hausner ratio is calculated with the relation Hr = V0 /Vf . This simple test has three major drawbacks. First, the result of the measurement depends on the operator. Indeed, the filling method influences the initial volume V0 . Secondly, the volume measurements by naked eyes induce strong errors on the results. Finally, with this simple method, we completely miss the compaction dynamics between the initial and the final measurements. GranuPaq instrument (Lumay et al., 2012; Traina et al., 2013) is an automated and improved tapped density measurement method based on recent fundamental research results (Lumay and Vandewalle, 2004, 2005; Richard et al., 2005; Lumay et al., 2006, 2009; Valverde and Castellanos, 2006). The behavior of the powder submitted to successive taps is analyzed with an automatized device. The Hausner ratios Hr, the initial density (0) and the final density (500) are measured precisely. Moreover, a dynamical parameter n1/2 and an extrapolation of the maximum density ∞ are extracted from compaction curves. Additional indexes can be used (Traina et al., 2013) but they are not presented in this paper. A compaction curve is a plot of the bulk density as a function of the tap number. The powder is placed in a metallic tube with a rigorous automated initialization process (Lumay et al., 2012). Afterward, a light hollow cylinder is placed on the top of the powder bed to keep the powder/air interface flat during the compaction process. The tube containing the powder sample rose up to a height of Z = 1 mm and performs free falls. The free fall height Z can be adjusted. The height h of the powder bed is measured automatically after each tap. From the height h, the volume V of the pile is computed. As the

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

315

Table 4 Main granulometric descriptors extracted from grain size distributions obtained with the laser diffraction method in dry method for air dispersive pressure at 1.2 bars. The sizes are expressed in ␮m. Std. is the standard deviation corresponding to the volume mean diameter (de Brouckere Mean Diameter) D[4/3]. Powder

d(0.1)

d(0.5)

d(0.9)

D[4/3]

Std.

T70 T80 T100

25.0 21.8 25.8

110.8 95.8 92.5

242.2 252.3 258.1

124.9 120.0 121.5

84.7 98.6 103.5

G70 G140

29.7 10.3

121.1 64.8

234.3 152.4

129.6 74.0

76.2 54.5

I70 I120 I230

90.5 87.3 58.6

167.1 132.1 97.5

285.9 198.9 152.2

178.1 138.4 101.0

77.4 43.5 38.9

F90 F100

65.5 41.6

134.1 129.6

234.4 247.7

142.3 138.8

66.2 76.9

powder mass m is known, the density  is evaluated and plotted after each tap. The density is the ratio between the mass m and the powder bed volume V. The measurements have been performed with 35 ml of powder subjected to 500 taps. The Hausner ratio Hr is related to the compaction ratio and is calculated by the equation Hr = (0)/(500), where (0) is the bulk density and (500) the tapped density computed after 500 taps. 2.2.4. Scanning electron microscopy Scanning electronic microscopy (Philips ESEM XL30 FEG, Eindhoven, NL) was used to characterize the particle size and morphology of the different lactose powders. Samples were deposited on carbon tapes. Sputtering deposition was done with gold target under argon atmosphere (Balzers, SCD004, Sputter coater). 2.2.5. Particle sizing The grain size distributions have been measured with the dry method with a laser diffraction particle size analyzer (Malvern, Mastersizer Sirocco 2000). The measurements have been made with different air injection pressures between 0.5 and 2 bar. 3. Results and discussion SEM micrographs of the different powders are presented in Figs. 2 and 3. For each powder sample, a micrograph showing the grains and a zoomed micrograph showing the grain surface are presented. Tablettose grains are agglomerates of primary particles. Granulac and Inhalac grains show angular edges due to the milling process. Finally, Flowlac grains have roughly spherical shapes related to the spray drying production process. We observe strong qualitative difference of both shape and surface roughness between the different families and also between the different powders inside a family. However, these differences are not discussed quantitatively in the present paper. The grain size distributions obtained by laser diffraction are presented in Fig. 4 and the main descriptors extracted from these distributions are summarized in Table 4. The presented distributions are corresponding to an air pressure of 1.2 bars. The main parameter selected to characterize the average grain size in the present study is the volume mean diameter (de Brouckere Mean Diameter) D[4/3]. For Granulac and Inhalac powders, the volume mean diameter D[4/3] significantly decreases when the number associated to the name increases (see Table 4). However, the differences between the powders inside both Tablettose and Flowlac families are less pronounced. We will discuss the link between the grain granulometric characteristics and the flowing properties hereafter.

Fig. 2. SEM micrographs of the Tablettose and Granulac powders.

In order to estimate the grain robustness, the granulometry has been measured with different air pressures. The measurement have been performed with 0.5, 0.8, 1.2, 1.6 and 2 bars. The volume mean diameter D[4/3] is plotted as a function of the air pressure in Fig. 5. For Tablettose powders, a clear decrease of the volume mean diameter D[4/3] is observed when the air pressure increases. This effect can be explained by the fact that Tablettose grains are agglomerates of smaller primary particles (see Fig. 2). The volume moment mean of Flowlac powders is also decreasing with the air pressure. At the opposite, Granulac grains are not broken by the air flow. The robustness of the Inhalac grains is intermediate with a slight decrease of the mean size when the air pressure increases. The results concerning lactose powders static properties obtained with the heap shape measurement method are presented in Table 5. All the measurements have been repeated three times and the standard deviation values are presented. These fast and straightforward measurements allow to perform a first classification of the samples as a function of the static cohesiveness. From low to high cohesiveness, we obtain roughly the following powder families classifying: Flowlac, Tablettose, Inhalac and Granulac. To get more advanced rheological information about the powder samples, a dynamic test like the rotating drum method is needed. The results related to the flowability obtained with GranuDrum are presented in Fig. 6 for the flowing angle ˛f and in Fig. 7 for

316

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

% volume

16 14 12 10 8 6 4 2 I 70

16 14 12 10 8 6 4 2 16 14 12 10 8 6 4 2

T 70

16 14 12 10 8 6 4 2 0.1

1

F 90

F 100

I 120

I 230

G 70

G 140

T 80

T 100

10

100

1000

10000

grain size (μm) Fig. 4. Size distributions of the 10 powder samples obtained with a dry laser diffraction method for air dispersive pressure at 1.2 bars.

Flowlac and Tablettose powders are characterized by both a low cohesive index  f and a low flowing angle ˛f . Therefore the flowing properties of these powders can be considered as good (see Table 3). Flowlac 90 flowability could be even considered as excellent. At the

300 250 200 150 100 50

Fig. 3. SEM micrographs of the Inhalac and Flowlac powders. 10 rpm

the dynamic cohesive index  f . Moreover, the flowing angle ˛f

and the cohesive index at the intermediate rotating speed of 10 rpm are showed in Table 6 with the first avalanche angle ˛a and the powder aeration (%ae ). These results show that the powders considered in the present study are covering a wide range of flowing properties. Some typical pictures of the powder flow inside the drum are showed in Fig. 1. Table 5 Results obtained from the heap shape analysis (GranuHeap instrument). The repose angle ˛r and the cohesive index  r are presented with the standard deviation (Std. dev.) evaluated from three measurements. Powder

˛r

Std. dev.

r

Std. dev.

T70 T80 T100

41.6 48.6 44.8

0.1 0.5 0.3

0.18 0.59 0.42

0.08 0.03 0.06

G70 G140

60.6 66.4

0.5 0.5

1.09 2.79

0.37 0.89

I70 I120 I230

39.4 47.1 53.6

0.5 0.3 0.5

0.61 0.44 0.58

0.08 0.06 0.07

F90 F100

38.6 40.1

0.7 0.5

0.20 0.15

0.04 0.02

D[4/3] (μm)

10 rpm f

300 250 200 150 100 50

I 70

300 250 200 150 100 50 300 250 200 150 100 50

T 70

0.5

F 90

F 100

I 120

I 230

G 70

G 140

T 80

T 100

1

1.5

2

Pressure (bar) Fig. 5. Volume mean diameter D[4/3] obtained with the laser diffraction granulometer in dry method as a function of the air pressure. The measurement have been performed with 0.5, 0.8, 1.2, 1.6 and 2 bars. Some points are slightly shifted horizontally in order to avoid symbols overlapping.

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

F 90

Table 6 Results obtained from the rotating drum analysis (GranuDrum instrument): flowing angle ˛f , dynamic cohesive index  f , first avalanche angle ˛a and powder aeration (%ae ).

F 100

60 50 40 I 70

I 120

I 230

60 50

10 rpm

10 rpm

Powder

˛f

f

˛a

T70 T 80 T100

37.7 41.7 42.6

12.1 23.0 14.6

42.4 38.0 43.0

0.8 5.8 3.8

G70 G140

51.2 57.9

34.7 53.6

59.7 68.4

1.6 35.2

I70 I120 I230

37.6 46.9 52.7

5.7 17.5 27.4

42.4 49.6 53.7

5.2 6.5 2.6

F90 F100

37.8 42.8

4.5 9.5

36.6 46.5

2.9 0.2

%ae

Flowing angle

f

(º)

317

40 G 70

G 140

60 50 40 T 70

T 80

T 100

60 50 40 2

4

6

8

10

12

14

16

18

20

Rotating Speed (RPM) Fig. 6. Flowing angle ˛f measured with the GranuDrum instrument as a function of the rotating speed. The powder families are plotted in different graphs to improve the clarity. The axis scales are identical to facilitate the comparisons. From top to bottom, the results obtained with Flowlac, Inhalac, Granulac and Tabletose powders are showed. Error bars are indicated for Granulac 140, Inhalac 70 and Inhalac 230.

Cohesive index

f

(AU)

60 50 40 30 20 10 I 70

60 50 40 30 20 10 60 50 40 30 20 10

T 70

60 50 40 30 20 10 2

4

6

F 90

F 100

I 120

I 230

G 70

G 140

T 80

T 100

8 10 12 14 16 Rotating Speed (RPM)

18

20

Fig. 7. Powder cohesiveness  f measured with the GranuDrum instrument the rotating drum as a function of the rotating speed. The powder families are plotted in different graphs to improve the clarity. The axis scales are identical to facilitate the comparisons. From top to bottom, the results obtained with Flowlac, Inhalac, Granulac and Tabletose powders are shown. Error bars are indicated for Granulac 140, Inhalac 70 and Inhalac 230.

opposite, the Granulac powders have higher cohesive index  f and flowing angle ˛f . Then, the general flowing properties of these powders are inferior. Concerning the flowability, the Inhalac powders are situated in between. At the exception of Tablettose powders family, the cohesive index always increase inside a family when the number associated to the powder increases. For example, the cohesion is higher in Inhalac 120 than in Inhalac 70 and the cohesion is higher in Inhalac 230 than in Inhalac 120. This number is generally associated to the size of the grains. The correlation between the flowability and the average grain size is analyzed hereafter. The flowing angle ˛f increases naturally with the rotating speed for all the samples at the exception of Granulac 140 powder. Moreover, a decrease of the cohesive index with the rotating speed is also observed for Granulac 140. Therefore, the flowability of the Granulac 140 is improved when the shear rate increases, corresponding to a shear-thinning behavior. The first avalanche angle ˛a is found to be correlated with the flowing angle ˛f . The powder aeration %ae during the flow are small at the exception of the Granulac 140 sample. Therefore, the shearthinning behavior of the Granulac 140 is certainly related to an aeration process during the flow. In order to check the reproducibility of the results, the measurements have been repeated three times for powders Inhalac 70, Inhalac 230 and Granulac 140. Inhalac 70 is one of the less cohesive powder while Granulac 140 is one of the more cohesive. The cohesiveness of Inhalac 230 is intermediate. Then, we have selected powders covering the whole range of physical characteristics. For these selected powders, the errors bars corresponding to the standard deviation are showed in the plots. The reproducibility is particularly good for the cohesive index  f . Indeed, error bars have a size comparable to the size of the symbols in the plot (see Fig. 7). The density of a powder bed and the ability of the powder bed to increase its density when submitted to vibrations are important informations when powders are used in production lines. Then, the bulk density (0), the tapped density (500) and an index related to the compaction dynamics are inescapable parameters. The compaction curves obtained with the GranuPaq instrument are presented in Fig. 8. The parameters extracted from these curves (initial density (0), tapped density (500), tap number to reach one half of the compaction process n1/2 , Hausner ratio Hr, and extrapolated optimal density (∞)) are summarized in Table 7. The results obtained in the present study are quite well correlated with the informations provided by the powder producer Meggle (see Table 1). The measurement has been repeated three times with the samples Granulac 140, Inhalac 70 and Inhalac 230. The results show that the measurements are reproducible and have high accuracy. Indeed, the repeated curves are not distinguishable in the plots of Fig. 8. This degree of reproducibility and of accuracy does not exist

318

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

3.0

1.5

(a)

0.80

1.4

σr

Hr

2.0 0.70

1.0

0.60 F 90

F 100

1.3 1.2

0.0

1.1 0 10 20 30 40 50 60

70

I 70

I 120

I 230

(c)

50 40

0.80

40

50

60

0.60

0.70

10 rpm

of dynamic cohesive index f

0.60 T 80

T 70 100

30

40

50

60

400

. (c) Repose angle ˛r measured with heap shape

method as a function of angle of first avalanche measured with rotating drum. (d) Repose angle ˛r measured with heap shape method as a function of flowing angle 10 rpm ˛f measured with rotating drum at 10 rpm. Squares, balls, triangles and inverted

T 100

200 300 tap number n

70

Fig. 9. Correlations between the results obtained with the different flowability measurement techniques (GranuDrum, GranuPaq and GranuHeap). (a) Static cohesive index  r measured with heap shape method as a function of dynamic cohesive index 10 rpm measured with rotating drum at 10 rpm. (b) Hausner ratio Hr as a function f

0.80

0

50

α f10rpm

α a(º) G 140

(d)

30 30

G 70

60

40

30

0.70

σ f10rpm

70

α r (◦ )

α r (◦ )

(g/ml) density

0.60

60

0 10 20 30 40 50 60

σ f10rpm

0.80 0.70

(b)

500

triangles are corresponding respectively to Granulac, Inhalac, Flowlac and Tablettose powders.

Fig. 8. Compaction curves obtained with the GranuPaq instrument. The powder families are plotted in different graphs to improve the clarity. The axis scales are identical to facilitate the comparisons. From top to bottom, the results obtained with Flowlac, Inhalac, Granulac and Tabletose powders are shown. The measure has been repeated three times with the samples Granulac 140, Inhalac 70 and Inhalac 230. As the measurements are reproducible, the repeated curves are not distinguishable.

with the classical tapped density measurement method described in the Pharmacopea. The density is related to all the grains properties (average size, size distribution span, shape, etc.) (Shenoy et al., 2015) and to the cohesive forces (van der Waals, electrostatic and capillary forces) (Lumay et al., 2009; Valverde and Castellanos, 2006) between the grains. Therefore, linking the density with the grains properties is a difficult task when the powder characteristics are varied. However, we will show hereafter that the Hausner ratio Hr is well correlated with the grain size and with the powder cohesiveness. The main correlations between the different flowability indexes are presented in Fig. 9. The static cohesive index  r extracted from the heap shape is found to be correlated with the dynamic cohesive 10 rpm index f measured with the rotating drum. The obtained correlation coefficient (covariance of the two variables divided by the Table 7 Results obtained with the density test (GranuPaq instrument): initial density (0), tapped density (500), tap number to reach one-half of the compaction process n1/2 , Hausner ratio Hr, and extrapolated optimal density (∞). The densities are expressed in g/ml. Powder

(0)

(500)

n1/2

Hr

(∞)

T70 T80 T100

0.553 0.638 0.561

0.642 0.782 0.661

27 26 31

1.162 1.226 1.177

0.679 0.839 0.704

G70 G140

0.683 0.557

0.876 0.820

25 34

1.282 1.473

0.946 0.922

I70 I120 I230

0.587 0.708 0.691

0.659 0.813 0.829

29 17 16

1.122 1.148 1.199

0.688 0.850 0.874

F90 F100

0.562 0.600

0.642 0.666

26 28

1.144 1.109

0.675 0.693

10 rpm

product of their standard deviations) is corr(r , f ) = 0.809. Therefore, when a set of samples have to be characterized, the straightforward heap shape measurement method is a good choice to perform a first screening. Afterward, the dynamic rotating drum method allow to obtain more informations about the powder rheological properties. As expected, the angle of repose ˛r is well correlated with the first avalanche angle ˛a with a correlation coefficient corr(˛r , ˛a ) = 0.806. Moreover, the angle of repose ˛r is also 10 rpm correlated with the flowing angle ˛f with a correlation coeffi10 rpm

) = 0.835. A less obvious correlation is found cient corr(˛r , ˛f between the Hausner ratio Hr and the dynamic cohesive index 10 rpm 10 rpm f . This correlation with a coefficient corr(Hr, f ) = 0.835 means that for a cohesive powder, the difference between the tapped density (500) and the bulk density (0) is higher. Therefore, the Hausner ratio Hr can be used to compare powders with different cohesiveness. Fig. 10 shows the relation between the main flowability indexes and the average grain size. The volume mean diameter D[4/3] has been selected to characterize the grain size. The cohesive 10 rpm index f , the angle of repose ˛r and the Hausner ratio Hr are decreasing when the grain size increases. The corresponding 10 rpm coefficients of correlation are respectively corr(f , D[4/3]) = −0.734, corr(˛r , D[4/3]) = −0.678 and corr(Hr, D[4/3]) = −0.707. When the grain size increases, the ratio between the weight of one grain and the cohesive forces (van der Waals, electrostatic and capillary forces) acting on this grain increases. Therefore, the influence of the cohesive forces on the powder bulk properties has the tendency to decrease as a function of the grain size (Lumay et al., 2012). In the present study, grains morphological properties are not analyzed quantitatively. However, we can obtain qualitative information from SEM micrographs. The considered lactose grain families can be sorted from the most angular shape to the most spherical shape. From SEM micrographs observation, the following classification from angular to spherical shape is obtained: Granulac, Inhalac, Tablettose and Flowlac. By considering a selected set of powders with comparable mean diameters, the effect of

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

319

powders can be considered as good. At the opposite, the Granulac powders have higher cohesive index  f and flowing angle ˛f . Then, the general flowing properties of these powders are inferior. However, Granulac 140 shows a shear-thinning behavior related to the aeration during the flow. Concerning the flowability, Inhalac powders are situated in between. This wide set of results concerning common lactose powders shows that high level measurement methods are necessary to determine precisely powders physical properties. Indeed, powders are complex materials with properties influenced by many parameters: grain sizes and shapes, surface roughness, surface chemical properties, electrostatic forces, van der Waals interactions and possibly capillary interactions. The correlations between the different flowability indexes have been analyzed. The informations obtained with the heap shape and the rotating drum methods are coherent. The Hausner ratio Hr is also found to be correlated with the powder cohesiveness. The link between the powder flowability and the mean grain size has been established. Indeed, when the average grain size decreases, the influence of the cohesive forces (van der Waals, electrostatic and capillary forces) on the powder bed behavior increases. Finally, the effect of grains shape on flowability has been analyzed qualitatively. Lactose powders made of angular grains have lower flowability than powders made of spherical grains. Acknowledgements Thanks to the Cat (microscopy center of the University of Liège) for the SEM micrographs. We thanks Meggle for giving free powder samples. Fig. 10. Correlations between the results obtained with the different flowability measurement techniques and the volume mean diameter D[4/3]. (a) Dynamic cohe10 rpm as a function of the volume mean diameter D[4/3]. (b) Repose angle sive index f

˛r as a function of the volume moment mean D[4/3]. (c) Hausner ratio Hr as a function of the volume mean diameter D[4/3]. Squares, balls, triangles and inverted triangles are corresponding respectively to Granulac, Inhalac, Flowlac and Tablettose powders. For powders with a volume mean diameter situated roughly between 120 and 140 ␮m, a vertical shaded off is shown from angular to round grains.

grains shape on flowing properties can be analyzed. In Fig. 10, a set of powders with a volume mean diameter D[4/3] situated roughly between 120 and 140 ␮m has been highlighted. Within this selected set of powders, the results are distributed vertically with angular shapes situated at the top and spherical shapes situated at the bottom. Therefore, angular shape leads to lower flowability. This effect is particularly pronounced when the repose angle ˛r is considered. Indeed, it is well known that grain shape modify drastically the angle of repose (Lumay et al., 2012). In addition to the angular shape, the higher fraction of fine grains observed in Granulac 70 powder (see Fig. 4) could also explain the higher value of 10 rpm f , ˛r and Hr obtained with this powder. 4. Conclusion Granulometry and flowability of 10 lactose powders have been analyzed. A first screening of the powder flowability has been performed with the straightforward heap shape measurement method (GranuHeap instrument). Afterward, the flowing behavior have been studied more deeply with the rotating drum method (GranuDrum instrument). Both measurement methods allow to measure significative differences between the powder samples. In particular, the influence of the cohesive forces (van der Waals, electrostatic and capillary forces) on the static and on the dynamic powder properties have been estimated. Flowlac and Tablettose powders are characterized by both a low cohesive index  f and a low flowing angle ˛f . Therefore the flowing properties of these

References Ashurst, I., Malton, A., Prime, D., Sumby, B., 2000. Latest advances in the development of dry powder inhalers. Pharmaceut. Sci. Technol. Today 3, 246–256. Braumann, A., Kraft, M., Mort, P.R., 2010. Parameter estimation in a multidimensional granulation model. Powder Technol. 197, 196–210. Chaudhuri, B., Alexandar, A.W.A., Faqih, A., Muzzio, F.J., Davis, C., Tomassone, M.S., 2006. Avalanching flow of cohesive powders. Powder Technol. 164, 13–21. de Gennes, P.G., 1999. Granular matter: a tentative view. Rev. Mod. Phys. 71, S374. de Ryck, A., Condotta, R., Dodds, J.A., 2005. Shape of a cohesive granular heap. Powder Technol. 157, 72. de Ryck, A., Zhu, H.P., Wu, S.M., Yu, A.B., Zulli, P., 2010. Numerical and theoretical investigation of the surface flows of granular materials on heaps. Powder Technol. 203, 125. European Pharmacopoeia 7.0, Powder Flow, p. 308 (Chapter 2.9.36). Fischer, R., Gondret, P., Rabaud, M., 2009. Transition by intermittency in granular matter: from discontinuous avalanches to continuous flow. Phys. Rev. Lett. 103, 128002. Freeman, R., 2007. Measuring the flow properties of consolidated, conditioned and aerated powders – a comparative study using a powder rheometer and a rotational shear cell. Powder Technol. 174, 25. Kaialy, W., Nokhodchi, A., 2015. Dry powder inhalers: physicochemical and aerosolization properties of several size-fractions of a promising alterative carrier, freeze-dried mannitol. Eur. J. Pharmaceut. Sci. 68, 56–67. Keleb, E.I., Vermeire, A., Vervaet, C., Remon, J.-P., 2004. Single-step granulation/tabletting of different grades of lactose: a comparison with high shear granulation and compression. Eur. J. Pharmaceut. Biopharmaceut. 58, 77–82. Kumar, A., Vercruysse, J., Bellandi, G., Gernaey, K.V., Vervaet, C., Remon, J.-P., De Beer, T., Nopens, I., 2014. Experimental investigation of granule size and shape dynamics in twin-screw granulation. Int. J. Pharmaceut. 475, 485–495. Lumay, G., Vandewalle, N., 2004. Compaction of anisotropic granular materials: experiments and simulations. Phys. Rev. E 70, 051314. Lumay, G., Vandewalle, N., 2005. Experimental study of granular compaction dynamics at different scales: grain mobility, hexagonal domains, and packing fraction. Phys. Rev. Lett. 95, 028002. Lumay, G., Vandewalle, N., 2010. Flow of magnetized grains in a rotating drum. Phys. Rev. E 82, 040301. Lumay, G., Vandewalle, N., Bodson, C., Delattre, L., Gerasimov, O., 2006. Linking compaction dynamics to the flow properties of powders. Appl. Phys. Lett. 89, 093505. Lumay, G., Dorbolo, S., Vandewalle, N., 2009. Compaction dynamics of a magnetized powder. Phys. Rev. E 80, 041302. Lumay, G., Boschini, F., Traina, K., Bontempi, S., Remy, J.-C., Cloots, R., Vandewall, N., 2012. Measuring the flowing properties of powders and grains. Powder Technol. 224, 19–27.

320

F. Boschini et al. / International Journal of Pharmaceutics 494 (2015) 312–320

Pirard, S.L., Lumay, G., Vandewalle, N., Pirard, J.-P., 2009. Motion of carbon nanotubes in a rotating drum: the dynamic angle of repose and a bed behavior diagram. Chem. Eng. J. 146, 143. Quintanilla, M.A.S., Valverde, J.M., Castellanos, A., 2006. The transitional behavior of avalanches in cohesive granular materials. J. Stat. Mech., P07015. Rajchenbach, J., 1990. Flow in powders: from discrete avalanches to continuous regime. Phys. Rev. Lett. 65, 2221. Richard, P., Nicodemi, M., Delannay, R., Ribière, P., Bideau, D., 2005. Slow relaxation and compaction of granular systems. Nat. Mater. 4, 121. Saw, H.Y., Davies, C.E., Jones, J.R., Paterson, A.H.J., 2014. Shear testing of lactose powders: the influence of consolidation stress and particle size on bulk density and estimated cohesion. Adv. Powder Technol. 25, 1164–1170. Schulze, D., 2011. Round Robin test on ring shear testers. Adv. Powder Technol. 22, 197–202. Shenoy, P., Viau, M., Tammel, K., Innings, F., Fitzpatrick, J., Ahrné, L., 2015. Effect of powder densities, particle size and shape on mixture quality of binary food powder mixtures. Powder Technol. 272, 165–172.

Taberlet, N., Richard, P., Hinch, E.J., 2006. S shape of a granular pile in a rotating drum. Phys. Rev. E 73, 050301. Tavares Cardoso, M.A., Talebi, M., Soares, P.A.M.H., Yurteri, C.U., van Ommen, J.R., 2011. Functionalization of lactose as a biological carrier for bovine serum albumin by electrospraying. Int. J. Pharmaceut. 414, 1–5. Tinke, A.P., Govoreanu, R., Weuts, I., Vanhoutte, K., De Smaele, D., 2009. A review of underlying fundamentals in a wet dispersion size analysis of powders. Powder Technol. 196, 102–114. Traina, K., Cloots, R., Bontempi, S., Lumay, G., Vandewalle, N., Boschini, F., 2013. Flow abilities of powders and granular materials evidenced from dynamical tap density measurement. Powder Technol. 235, 842–852. Valverde, J.M., Castellanos, A., 2006. Random loose packing of cohesive granular materials. Europhys. Lett. 75, 985. Warnke, G., Konig, D., Bauer, K.H., Brandl, M., 1996. Aerodynamic separation of a salbutamol dry powder inhalation: deposition of drug and excipient in a multistage liquid impinger. Pharmaceut. Pharmacol. Lett. 6, 64–67.