1

Investigation of internal structure of fine granules by microtomography using synchrotron

X-ray radiation

Shuji Noguchi a, Ryusuke Kajihara a, Yasunori Iwao a, Yukari Fujinami a, Yoshio Suzuki b,

Yasuko Terada b, Kentaro Uesugi b, Keiko Miura b and Shigeru Itai a *5

a School of Pharmaceutical Sciences, University of Shizuoka, 52-1 Yada, Suruga-ku, Shizuoka

422-8526, Japan

b Japan Synchrotron Radiation Research Institute, 1-1-1 Kouto, Sayo-gun, Hyogo 679-5198,

Japan10

* Corresponding author. E-mail: s-itai@u-shizuoka-ken.ac.jp; Tel: +81-54-246-5614; Fax:

+81-54-264-5615.

Abbreviations: APAP, acetaminophen; AMCE, aminoalkyl methacrylate copolymer E; BHX,

bromhexine hydrochloride; CT, computed tomography; DCPD, dibasic calcium phosphate 15

dihydrate; GM, glyceryl monostearate; HPC, hydroxypropyl cellulose; LAC, linear attenuation

coefficient; MC-wax, microcrystalline wax; SEM, scanning electron microscopy

2

Abstract20

Computed tomography (CT) using synchrotron X-ray radiation was evaluated as a

non-destructive structural analysis method for fine granules. Two kinds of granules have been

investigated: a bromhexine hydrochloride (BHX)-layered Celphere® CP-102 granule coated

with pH-sensitive polymer Kollicoat® Smartseal 30-D, and a wax-matrix granule constructed

from acetaminophen (APAP), dibasic calcium phosphate dehydrate, and aminoalkyl 25

methacrylate copolymer E (AMCE) manufactured by melt granulation. The diameters of both

granules were 200–300 μm. CT analysis of CP-102 granule could visualize the laminar

structures of BHX and Kollicoat layers, and also visualize the high talc-content regions in the

Kollicoat layer that could not be detected by scanning electron microscopy. Moreover, CT

analysis using X-ray energies above the absorption edge of Br specifically enhanced the contrast 30

in the BHX layer. As for granules manufactured by melt granulation, CT analysis revealed that

they had a small inner void space due to a uniform distribution of APAP and other excipients.

The distribution of AMCE revealed by CT analysis was also found to involve in the differences

of drug dissolution from the granules as described previously. These observations demonstrate

that CT analysis using synchrotron X-ray radiation is a powerful method for the detailed internal 35

structure analysis of fine granules.

Keywords: Computed tomography, synchrotron X-ray radiation, polymer-coated granule, melt

granulation.

40

3

1. Introduction

As the characteristics of a drug release from the granular formulation, for example

pH-dependent release of drugs, are closely related to the structure, structural information on

granules containing drugs is indispensable to elucidate the drug release mechanism and to

design new or more potent formulations. The structural information of granules containing 45

drugs includes the distribution of the drug or excipient particles, presence of void spaces,

thickness of the drug or polymer layers, and shape or even the roughness of the surface, and

various analytical methods have been applied to obtain these structures in detail. Scanning

electron microscopy (SEM) is the most widely used method to observe the surface structures of

granules. Raman spectroscopy has also been successfully used to reveal the distribution of drugs 50

and excipients (Gordon and McGoverin, 2010). However, these methods reveal only the

structural information about the surface or inner region just below the surface of the granules,

and in order to observe internal structures, samples must be cut or destroyed. In an investigation

by SEM, samples are placed under vacuum and deposited with a heavy metal such as platinum,

which can affect the surface structure of the sample. Although magnetic resonance imaging can55

reveal the internal structure of a drug formulation without destruction, it has been limited to

apply for tablets (Broadbent et al., 2010) owing to its low spatial resolution.

X-ray computed tomography (CT) can elucidate the three-dimensional structure of a

formulation non-destructively owing to the high penetration of X-ray, and has been applied to

the structural analysis of tablets (Li et al., 2012) and granules (Crean et al., 2010) with sizes60

ranging from centimeters to millimeters using an in-house X-ray generator. Recently,

high-quality X-ray micro CT analysis on sub-mm specimens has become possible with

submicrometer spatial resolution. This has been used on mineral samples from asteroids

(Tsuchiyama et al., 2011) using synchrotron X-ray radiation, which is extremely bright and

4

highly parallel and has a tunable energy. Although the synchrotron X-ray CT analysis of tablets65

(Laity and Cameron, 2010) and granules for tableting (Morita and Yamahara, 2006) has been

performed, it has not been applied to a structural analysis of fine granules such as

polymer-coated granules manufactured by agitation fluidized bed coating processes and

wax-matrix granules manufactured by melt granulation. These fine granules are thought to have

a layer structure of micro-meter order thickness under the surface or contain particles of70

micro-meter order diameters in their inside. Non-destructive and three-dimensional visualization

of these structural features would be best performed by synchrotron X-ray CT of

sub-micrometer spatial resolution. Fine granular formulations give a good mouth feel upon oral

administration when their diameters are 200 μm or less. Since fine granules are used in

orally-disintegrating tablets, the demands on the production of such drug-containing fine 75

granules are growing. We report here on the use of micro CT using synchrotron X-ray radiation

as a method for the non-destructive analysis of the internal structure of pharmaceutical granules

with 200–300 μm diameters.

2. Material and Methods80

2.1 Materials

Celphere® CP-102 was kindly provided by Asahi Kasei (Tokyo, Japan), acetaminophen

(APAP) by Iwaki Pharmaceutical Co., Ltd. (Shizuoka, Japan), Kollicoat® Smartseal 30D by

BASF Japan Ltd. (Tokyo, Japan), dibasic calcium phosphate dihydrate (DCPD) by Kimura

Sangyo Co., Ltd. (Tokyo, Japan), and microcrystalline wax (MC-wax) by Nippon Seiro Co., Ltd. 85

(Tokyo, Japan). Bromhexine hydrochloride (BHX) was purchased from Shiratori

Pharmaceuticals Inc. (Chiba, Japan), aminoalkyl methacrylate copolymer E (AMCE) from

Röhm Degussa (Darmstadt, Germany), and glyceryl monostearate (GM) from Taiyo chemical

5

Industry Co., Ltd. (Saitama, Japan). All reagents used were of the highest grade available from

the commercial source.90

2.2 Preparation of the fine granules

2.2.1 Polymer-coated fine granules

Eighty grams of BHX, 16 g of hydroxypropyl cellulose (HPC), and 4 g of polyethylene

glycol 6000 were mixed with 420 g of water and 280 g of ethanol. The mixture was 95

homogenized three times with a microfluidizer (M110-E/H, Powrex, Hyogo, Japan) at 175 MPa.

The homogenized solution was then layered onto 700 g of Celphere® CP-102 by side-spraying

in a fluidized-bed granulator MP-101 (Powrex, Hyogo, Japan). Layering conditions were as

follows: inlet air temperature, 50–55°C; air flow, 0.55 m3/min; product temperature, 28–29°C;

atomizing air pressure, 0.5 MPa; spray rate, 6 g/min. After spraying, the drug-layered granules 100

were dried under a flow of air at 28°C for 10 min. The granules were sieved through a 210-μm

mesh.

For a polymer coating, the solution containing 200 g of Kollicoat® Smartseal 30D, 9 g of

triethyl citrate, 48 g of talc and 343 g of water was mixed with vigorous stirring for 10 min. Five

hundred grams of this Kollicoat® solution was sprayed onto the 500 g of the BHX-layered 105

granules by fluidized-bed granulator side-spraying under the same conditions as above except

that the spray rate was decreased to 4 g/min. After Kollicoat coating, a solution containing 12 g

of D-mannitol, 0.08 g of HPC-L, and 88 g of water was then sprayed under the same conditions

to prevent the aggregation of the polymer-coated granules, and they were dried in a flow of air

at 28°C for 10 min. The granules thus obtained were sieved through a 250-μm mesh.110

2.2.2 Granules with pH-dependent release by melt granulation

6

Granules for pH-dependent release were prepared by the method of Shiino et al. (2012). In

short, a mixture of APAP and DCPD was processed in a sample mill (Tl-300, Cosmic

Mechanical Technology Co., Ltd., Fukushima, Japan) and sieved through a 149-μm mesh. One 115

hundred eleven grams of an APAP/DCPD mixture, 15 g of AMCE, 12 g of GM, and 12 g of

MC-wax were put into a high-shear mixer (MECHANOMiLL, Okada Seiko Co., Ltd., Tokyo,

Japan), and the melt granulation was performed with an impeller speed of 1200 rpm and a

temperature of a jacket rubber heater set to 85°C. When the temperature of the mixture reached

to 75°C, heating was stopped, and mixing was continued for 2 min at a reduced impeller speed 120

of 400 rpm. After the product granules were cooled by spreading in a metal tray for 4 min,

additional granulation was performed for 1 min at an impeller speed of 400 rpm at room

temperature.

2.3 Synchrotron X-ray CT measurement125

The granules were placed in Lindemann glass capillaries 0.2–0.3 mm in diameter. X-ray CT

measurements were performed using a micro-CT instrument (Uesugi et al., 2012; Suzuki et al.,

2011) installed at the undulator beam line BL37XU of SPring-8 (Hyogo, Japan). The X-ray

energy was set to 8 keV, and flux density was approximately 6×1012 photons/s/mm2. In the case

of the polymer-coated granules, CT measurements were also made at 15 keV X-ray. The flux 130

density at 15 keV was estimated to be approximately 60% of that at 8 keV, based on the

sensitivity and the scintillator thickness of the area detector. Nine hundred transmission X-ray

images, in a parallel projection geometry, were recorded in 0.2° steps with continuous rotation

of the sample (on-the-fly scan mode). During the measurement, the sample was irradiated with

X-ray continuously, and the exposure time for each projection was 150 ms. Sample to area 135

detector distance was 2 mm. CT measurement of one sample was accomplished within 5 min.

7

All the CT measurements were performed at room temperature. Tomographic reconstruction

was performed by the convolution back projection method using the CBP package (Uesugi et al.,

2004), and 1920 × 1920 pixel cross-sectional images were obtained. One pixel is equivalent to

0.444 μm × 0.444 μm. 3-D analyses of the cross-sectional images were performed using SLICE140

package (Nakano et al., 2006) and ImageJ (Schneider et al., 2012). X-ray linear attenuation

coefficients (LAC) values between 0 and 70 are shown in 8 bit grayscale in the images of

polymer-coated granules, with LAC value 70 and higher as white. In the case of wax-matrix

granules by melt granulation, LAC values between 0 and 130 are shown in 8 bit grayscale.

LACs of drugs and excipients were calculated using the software MU_3 (Kato, 2011; Hubbell 145

and Seltzer, 1996). The LAC values as a function of X-ray energy are shown in Fig. 1.

2.4 Scanning electron microscopy

The surface structure of the polymer-coated fine granules were morphologically assessed by

SEM (Model: JSM-5310LV; JEOL Ltd., Tokyo, Japan). The samples were placed on150

double-sided adhesive tape and were sputter-coated with platinum under vacuum prior to

imaging.

3. Results and Discussion

3.1. Structure of polymer-coated granules155

In the cross-sectional image of the polymer coated granule, laminar structures of BHX with

higher LAC values and outer Kollicoat layers are clearly observed at the surface of the cellulose

core particle CP-102, as shown in Fig. 2a. The thicknesses of the BHX layer and Kollicoat

coatings were approximately 2–5 μm and 5–10 μm, respectively. Although the mannitol was

sprayed onto the granules over the Kollicoat layer, the mannitol layer cannot be recognized, 160

8

possibly because the amount of mannitol was only one tenth the amount of the Kollicoat. The

boundary between the cellulose core particle and BHX layer is somewhat smeared. This is

probably because the BHX-dispersed solution containing ethanol might soak into the cellulose

core when the solution was sprayed. The Kollicoat layer contains regions with high LAC values,

which indicates they consist of talc (Fig. 1). They are 1–2 μm in thickness and 5–10 μm in 165

diameter. The presence of these high talc-containing regions in the Kollicoat layer indicate that

the talc was not completely dispersed in the Kollicoat mixture solution in spite of the vigorous

stirring and sprayed at high pressure. Such regions with high talc content are exposed to the

surface of the granules, as shown in Fig. 2b. Since talc is highly wettable, the regions might

disintegrate rapidly when the granules are in contact with outer solution, resulting in the 170

formation of regions with a thin polymer-coating or even pores, especially where the polymer

layer is thin. These would impede the function of the polymer layers. The presence of the high

talc regions cannot be detected in SEM image (Fig. 2c), because there is little difference in the

surface asperity of a talc-exposed region and a Kollicoat-layered region. This demonstrates

X-ray CT analysis might predominate over SEM even in the analysis of the surface structure of 175

fine granule.

Talc in Kollicoat layer and BHX can be distinguished based on contrast: their X-ray LAC

values at 8 keV are 75.5 cm-1 for BHX and 85.6 cm-1 for talc. The difference in the contrast can

be enhanced by doing the X-ray CT measurement at 15 keV, above the K absorption edge of

bromine (13.473 keV). At 15 keV, the LAC of BHX was 73.5 cm-1, whereas that of talc has a 180

much smaller value of 13.7 cm-1. In the cross-sectional image obtained using 15 keV X-ray, the

BHX layer is recognized as a high LAC values region as in 8-keV cross-sectional images,

whereas talc in Kollicoat layer is dark, in accordance with its lower LAC values (Fig. 2d). This

demonstrates that by setting the X-ray energies above and below the absorption edge of the

9

element of interest, X-ray CT using synchrotron radiation can visualize the internal distribution 185

of drugs or excipients containing a specific element in fine granules.

The color of sample became a pale-brown after two-successive CT measurement (Fig. 3a).

This might imply that some part of the material in the granule had decomposed by the intense

synchrotron X-ray radiation (Borek et al., 2007; Ennifar et al., 2002; Ravelli and Garman, 2006).

During the CT measurement at 8 keV, the sample and the glass capillary are exposed to 190

approximately 5 × 1013 photons, across a diameter of 250 μm. Figs. 3b and c show the

cross-sectional and surface images of the same formulation for two successive CT

measurements. X-ray would be absorbed more in materials with high LAC values, that is, the

layers containing BHX or talc. However, little structural differences are found in these layers or

other part of the granule. No discernible alternation of the surface structure was also observed195

between SEM images of granules that had or had not been subjected to CT measurement (Figs.

2c and e). These images indicate that any structural changes caused by radiation damage are

negligible, although the color change does indicate some changes. If the structural changes were

caused by intense synchrotron X-ray during the CT measurements, the cross-sectional images

would be distorted (Barett & Keat, 2004). No distortion was observed in the cross-sectional 200

images of the granules shown in Fig. 3. This also indicates that the structural change during CT

measurement is negligible. Since the X-ray flux density and LAC values of any excipient other

than BHX at 15 keV is lower than that at 8 keV, any radiation damage during CT measurement

using 15 keV X-ray can also be assumed to be negligible.

205

3.2. Structure of granule with pH-dependent release by melt granulation

Images of a granule prepared by melt granulation are shown in Fig. 4. In the cross-sectional

image (Fig. 4a), the granule can be divided into three groups based on the LAC values.

10

Sharp-edged regions with the highest LAC values are DCPD particles. DCPD has a highest

calculated LAC value, 140 cm-1, among the ingredients of this formulation. The maximum 210

diameter of the DCPD particles is approximately 70 μm, and fine particles with sub-micrometer

diameter can be also recognized. Between the DCPD particles, regions that have either

intermediate or low LAC values are present. The former may be GM and MC-wax, and the

latter, indicated by arrows A and B in Fig. 4a, may be APAP and AMCE. Void spaces are rare

inside the formulation, indicating that highly dense granules could be manufactured by this melt 215

granulation method. GM and MC-wax are indistinguishable, because the difference in their

LAC values is small owing to their similar elemental composition and would be thoroughly

mixed together while melted during the melt granulation process. Fine particles of DCPD with

diameters less than the pixel size in the cross-sectional image (0.444 μm) may be well-dispersed

in the melted waxes during the melt granulation process, giving a LAC value higher than the 220

pure wax. APAP or AMCE particles does not melt throughout the granulation process and have

small pores or cracks where fine particles of DCPD can penetrate, which leaves the LAC values

unchanged. This could explain why particles of AMCE and APAP are observed in the lower

LAC value regions more than other excipients in the cross-sectional images of this granule.

Although there is little contrast difference between AMCE and APAP, they can be distinguished225

based on their morphology. Particles with sharp edges, as indicated by arrow A in Fig. 4a,

should be a hard APAP crystal, and the particle deformed plastically by DCPD, as indicated by

arrow B in Fig. 4a, should be the softer AMCE. Particles of DCPD, APAP, and AMCE are

exposed on the surface of the granule (Fig. 4b–d). Indentations are found sporadically in the

wax regions as indicated by arrow C in Fig. 4a, and are thought to be formed when the hard 230

particles of DCPD or APAP slipped off the surface owing to collision or friction with the other

granules during the mixing in the high-shear mixer.

11

Previously, when these granules were used in the dissolution test (paddle method listed in the

Japanese Pharmacopoeia XVI) at pH 4.0 (0.1 M acetate buffer) and 37.0±0.5°C for 120

minutes, large cracks form on the surface of the granules as observed by SEM, which are 20–50235

μm in width (Shiino et al., 2012). These cracks formed on the dissolution of AMCE, as no crack

formation is observed in a granule without AMCE. This is consistent with the structure of the

granule as revealed by CT analysis, because cracks of that size would form when the AMCE

and APAP dissolves and DCPD particles come unstuck from the granule. The structure revealed

by CT analysis also suggests that the particle size of AMCE determines the morphology of the 240

granules upon dissolution. When the granules are prepared using AMCE with the diameter

smaller than those of APAP and DCPD, not large cracks but smaller pores might form on the

surface upon dissolution at pH 4.0, which would increase the surface area and result in an

enhanced pH dependence of the granules.

245

4. Conclusion

We have applied CT using synchrotron X-ray radiation for the structural analysis of two kinds

of fine granules containing drugs. The structural information obtained by CT analyses revealed

the layer structures and the drug and excipients distribution inside the granules as well as on the

surface. Drugs, containing bromine, could be clearly visualized in the cross-sectional image by 250

CT measurement using an X-ray energy above its absorption edge. This is possible only using

wavelength-tunable synchrotron X-ray radiation. This study demonstrates the high potency of

CT using synchrotron radiation to obtain structural information on fine granules, which will

improve their functionality and, moreover, aid the design of new fine granular formulations.

255

Acknowledgment

12

The synchrotron radiation experiments at BL37XU were performed with the approval of the

Japan Synchrotron Radiation Research Institute (JASRI; Proposal No. 2012A1670).

260

13

References

Barrett, J.F., Keat, N., 2004. Artifacts in CT: recognition and avoidance. Radiographics, 24,

1679–1691.

Beevers, C.A., 1958. The crystal structure of dicalcium phosphate dihydrate, CaHPO42H2O.

Acta Crystallogr. 11, 273–277.265

Borek, D., Ginell, S.L., Cymborowski, M., Minor, W. Otwinowski, Z., 2006. The many faces of

radiation-induced changes. J. Synchrotron Rad. 14, 24–33.

Botez, C.E., Stephens, P.W., 2003. Crystal structure of anhydrous δ-D-mannitol. Powder Diffr.

18, 214–218.

Broadbent, A. L., Fell, R. J., Codd, S. L., Kim, A. L., Konagurthu, S., Koehler-King, D. G., 270

Seymour, J. D., 2010. Magnetic resonance imaging and relaxometry to study water

transport mechanisms in a commercially available gastrointestinal therapeutic system

(GITS) tablet. Int. J. Pharm. 397, 27–35.

Crean, B., Parker, A., Le Roux, D., Perkins, M., Luk, S., Banks, S., Melia, C., Roberts, C., 2010.

Elucidation of the internal physical and chemical microstructure of pharmaceutical 275

granules using X-ray micro-computed tomography, Raman microscopy and infrared

spectroscopy. Eur. J. Pharm. Biopharm. 76, 498–506.

Ennifar, E., Carpentier, P., Ferrer, J.-L, Walter, P., Dumas, P., 2002. X-ray-induced

debromination of nucleic acids at the Br K absorption edge and implications for MAD

phasing. Acta Crystallogr. D58, 1262–1268.280

Gordon, K.C., McGoverin, C.M., 2010. Raman mapping of pharmaceuticals. Int. J. Pharm. 417,

151–162.

Haisa, M., Kashino, S., Kawai, R., Maeda, H., 1976. The monoclinic form of

p-hydroxyacetamide. Acta Crystallogr. B32, 1283.

14

Hubbell, J.H., Seltzer, S.M., 1996. Physical Reference Data. NISTIR 5632.285

Kato, H., 2011. http://www.fujita-hu.ac.jp/~hid-kato/freesoft.html

Koo, C.H., Jung, Y.J., Lee, S.W., 1984. The crystal and molecular structure of bromhexineHCl.

Arch. Pharm. Res. 7, 115–120.

Laity, P.R., Cameron, R.E., 2010. Synchrotron X-ray microtomograhpic study of tablet swelling.

Eur. J. Pharm. Biopharm. 75, 263–276.290

Li, H., Yin, X., Ji, J., Sun, L., Shao, Q., York, P., Xiao, T., He, Y., Zhang, J., 2012.

Microstructural investigation to the controlled release kinetics of monolith osmotic pump

tablets via synchrotron radiation X-ray microtomography. Int. J. Pharm. 427, 270–275.

Morita, T. and Yamahara, H., 2006. Three-dimensional imaging and analysis of granules by high

resolution X-ray computed microtomography at the third-generation synchrotron 295

radiation facility, SPring-8. Pharm. Tech. Japan, 22, 2533–2539.

Nakano, T., Tsuchiyama, A., Uesugi, K., Uesugi, M., Shinohara, K., 2006. "Slice" – Software

for basic 3-D analysis –. http://www-bl20.spring8.or.jp/slice/

O'Sullivan, A.C., 1997. Cellulose: the structure slowly unravels. Cellulose 4, 173–207.

Quillin, M.L., Matthews, B.W., 2000. Accurate calculation of the density of proteins. Acta 300

Crystallogr. D56, 791–794.

Rayner, J.H., Brown, G., 1972. The crystal structure of talc. Clay Clay Miner. 21, 103–114.

Ravelli, R.B.G., Garman, E.F., 2006. Radiation damage in macromolecular cyocrystallography,

Curr. Opin. Struc. Biol.16, 624–629.

Schneider, C.A., Rasband, W.S., Eliceiri, K.W., 2012. NIH Image to ImageJ: 25 years of image305

analysis. Nature Methods 9, 671–675.

Shiino, K., Iwao, Y., Fujinami, Y., Itai, S., 2012. Preparation and evaluation of granules with

pH-dependent release by melt granulation. Int. J. Pharm. 431, 70–77.

15

Suzuki, Y., Takeuchi, A., Uesugi, K., 2011. Current status of microtomography with synchrotron

radiation X-ray source. J. Vac. Soc. Jpn. 54, 47–55.310

Tsuchiyama, A., Uesugi, M., Matsushima, T., Michikami, T., et al., 2011. Three-dimensional

structure of Hayabusa samples: Origin and evolution of Itokawa regolith. Science, 333,

1125–1128.

Uesugi, K., 2004. http://www-bl20.spring8.or.jp/xct/index.html

Uesugi, K., Hoshino, M., Takeuchi, A., Suzuki, Y., Yagi, N., 2012. Development of fast and high 315

throughput tomography using CMOS image detector at SPring-8. SPIE Proceedings, in

press.

16

Figure captions

Figure 1. Calculated LAC values of the drugs and excipients as a function of X-ray energy. In 320

the calculation of LAC, the following densities were used: APAP 1.30 g/cm3 (Haisa et al, 1976),

BHX 1.62 g/cm3 (Koo et al., 1984), crystalline cellulose 1.59 g/cm3 (O'Sullivan, A.C., 1997),

talc 2.76 g/cm3 (Rayner and Brown, 1972), D-mannitol (β-form) 1.39 g/cm3 (Botez and

Stephens, 2003), GM 0.97 g/cm3 and DCPD 2.32 g/cm3 (Beevers, 1958). As the density of

AMCE is not reported, the average calculated density of biopolymer protein, 1.47 g/cm3325

(Quillin and Matthews, 2000), was tentatively used. X-ray energies used in the CT

measurements are indicated by vertical dotted lines.

Figure 2. Images of polymer-coated granules. (a) Cross-sectional and (b) surface images by CT

measurement at 8 keV. BHX layer, Kollicoat layer and some of the high talc-content regions are 330

indicated by arrows. Images of capillary were removed for clarity. (c) Surface image by SEM.

(d) Cross-sectional images by CT measurement at 15 keV. (a)–(d) are the same granule. (e)

SEM image of a granule not subjected to CT measurement.

Figure 3. (a) Microspcopic images of polymer-coated granules before CT measurement (left) 335

and after two-successive CT measurements (right). (b) Cross-sectional surface image by 1st CT

measurement at 8 keV. (c) is repeats of the same CT measurements. The edge of the trimmed

regions are shown with white lines for clarity. (a)–(c) are the same granules. The small rods

observed at the surface of the granule are the fragments of glass that were already present in the

capillary.340

Figure 4. Images of a granule manufactured by melt granulation. (a) Cross-sectional and (b)–(d)

17

surface images by CT measurement at 8 keV. The position of cross-sectional image (a) is shown

by dotted arrows in (c) and (d).

Figure 1

50 μm

50 μm

Figure 2

Kollicoatlayer

BHX layer

Talc

(b)

(a)

Talc

(c)

(d)

Figure 2

50 μmBHX layer

Figure 2

(e)

Figure 3

(a)

0.5 mm

50 μm

Figure 3

(c)

(b)

50 μm

Figure 4

(b)

100 μm

A

B

C

100 μm

(a)

y

z

x

y

(a)

100 μmx

z

(c)

100 μm

x

y

Figure 4

(d)

(a)