Pharmaceutical Technology DivisionDepartment of Pharmacy

University of HelsinkiFinland

Rheological properties and the state of water of

microcrystalline cellulose and silicified microcrystalline

cellulose wet masses

by

Pirjo Luukkonen

Academic Dissertation

To be presented, with the permission ofthe Faculty of Science of the University of Helsinki,

for public criticism in Auditorium 1 at Viikki Infocentre (Viikinkaari 11A)on June 16th, 2001, at 12 noon

Helsinki 2001

Supervisors: Professor Jouko YliruusiDivision of Pharmaceutical TechnologyDepartment of PharmacyUniversity of HelsinkiFinland

Docent Leena HellénOrion PharmaFinland

Reviewers: Docent Anne JuppoAstraZeneca R&D MölndalSweden

Dr. Kirsti SaarnivaaraOrion PharmaFinland

Opponent: Professor Hans LeuenbergerDepartment of Pharmaceutical TechnologyUniversity of BaselSwitzerland

© Pirjo Luukkonen 2001ISBN 952-10-0022-8ISBN 952-10-0023-6 (pdf, http://ethesis.helsinki.fi/)ISSN 1239-9469

YliopistopainoHelsinki 2001Finland

To my parents, Laura and Seppo

ABSTRACT

Luukkonen, P., 2001. Rheological properties and the state of water of microcrystalline cellulose and

silicified microcrystalline cellulose wet masses.

Dissertationes Biocentri Viikki Universitatis Helsingiensis 9/2001, 58 pp.

ISBN 952-10-0022-8 ISBN 952-10-0023-6 (pdf) ISSN 1239-9469

The aim of the present study was to investigate the rheological properties and thestate of water of microcrystalline cellulose (MCC) and silicified microcrystalline cellulose(SMCC) wet masses. Furthermore, the purpose was to understand the solid-waterinteractions and to compare the characterisation methods available for wet powdermasses.

The rheological properties of wet cellulose masses were studied at six differentwater contents (0.8-1.5 g g-1) by using mixer torque rheometer (MTR), powder rheometer(PR) and capillary rheometer. The state of water in the SMCC wet masses was studied atwide water range (0.05-1.5 g g-1) by using near infrared (NIR) spectroscopy anddifferential scanning calorimetry (DSC) measurements (thermoporosimetry). In addition,the physical properties of dry cellulose powders were characterised.

The wet masses exhibited similar behaviour in the rheometers based on thetorque measurements (MTR and PR). The cohesiveness of the wet masses and, thus, thetorque values increased with the increasing water content until the wet mass wasoverwetted. The capillary rheometer, instead, measured the standard rheologicalproperties of the materials. The rheological properties of MCC and SMCC wet masseswere slightly different and changed with the water content. Even though the SMCCgrade proved to be more elastic than the MCC grades, the cellulose wet masses could notbe differentiated with capillary rheometer as clearly as with the torque measurements.

NIR spectroscopy could be used to study the liquid retention capacity ofdifferent materials. The method was also able to distinguish the hydrate water from theother energetic states of water. The results could be correlated to the maximum torquevalues obtained in the MTR measurements, which indicates that the NIR spectroscopycould be used for the end-point detection in the real time processes. Thermoporosimetrywas able to differentiate four distinct fractions of water in MCC/SMCC wet masses.Furthermore, with this method it was possible to determine the pore structure of wetcellulose samples. The pore volume results were consistent with the NIR spectroscopyresults, which indicates that the changes in the state of water can be seen in a similarmanner with both the methods. Consequently, the methods used in this study gavetogether extensive information about the properties and behaviour of the wet cellulosemasses.

i

TABLE OF CONTENTSTable of contents _____________________________________________________ i

Acknowledgements __________________________________________________ iv

List of original publications ___________________________________________ vi

List of abbreviations ________________________________________________ vii

1. Introduction ___________________________________________ 1

2. Theory________________________________________________ 3

2.1 Water-solid interactions _________________________________________3

2.1.1 Structure of water _______________________________________________________3

2.1.2 Water sorption by solids __________________________________________________3

2.1.3 Wet agglomeration ______________________________________________________5

2.2 Special characteristics of cellulose with water________________________8

2.2.1 Structure of cellulose_____________________________________________________8

2.2.2 Water sorption by cellulose ________________________________________________9

2.2.3 Swelling______________________________________________________________11

2.2.4 Hornification__________________________________________________________11

2.2.5 Behaviour of MCC/SMCC in extrusion-spheronisation__________________________12

2.3 Wet mass characterisation_______________________________________ 13

2.3.1 Torque methods _______________________________________________________13

2.3.2 Rheological methods ____________________________________________________15

2.3.3 Thermal methods ______________________________________________________18

2.3.4 Molecular methods _____________________________________________________19

3. Aims of the study ______________________________________ 22

4. Experimental _________________________________________ 23

4.1 Materials _____________________________________________________23

ii

4.2 Characterisation of cellulose powders (I) __________________________23

4.3 Preparation of wet masses_______________________________________24

4.4 Mixer torque rheometry (I, II, IV) ________________________________24

4.5 Powder rheometry (II)__________________________________________24

4.6 Capillary rheometry (III)________________________________________25

4.7 Near infrared spectroscopy (IV)__________________________________25

4.8 Thermoporosimetry and solute exclusion (V)_______________________25

4.8.1 Isothermal step melting procedure _________________________________________26

4.8.2 Total bound water (TBW) measurements ____________________________________26

4.8.3 Solute exclusion technique________________________________________________26

5. Results and discussion _________________________________ 28

5.1 Torque measurements (I, II) ____________________________________28

5.1.1 Mixing kinetics ________________________________________________________28

5.1.2 Effect of mixing time ___________________________________________________29

5.1.3 Effect of shear forces ___________________________________________________30

5.1.4 Evaluation of methods __________________________________________________31

5.2 Rheological properties obtained by capillary rheometry (III) __________ 31

5.2.1 Flow curves___________________________________________________________31

5.2.2 Rheological models _____________________________________________________31

5.2.3 Elastic properties_______________________________________________________32

5.2.4 Evaluation of capillary rheometry __________________________________________32

5.3 Near infrared spectroscopy (IV)__________________________________33

5.3.1 Absorbance spectra _____________________________________________________33

5.3.2 Second derivative spectra ________________________________________________35

5.3.3 Evaluation of NIR spectroscopy ___________________________________________36

5.4 Thermoporosimetry (V) ________________________________________36

iii

5.4.1 Different water fractions in SMCC wet masses ________________________________36

5.4.2 Effect of granulation ____________________________________________________37

5.4.3 Evaluation of thermoporosimetry __________________________________________38

5.5 The difference between cellulose grades ___________________________39

5.5.1 SMCC vs. MCC grades __________________________________________________39

5.5.2 MCC grades __________________________________________________________42

5.6 Comparison of different methods_________________________________43

5.6.1 Capillary rheometry vs. MTR______________________________________________43

5.6.2 NIR vs. MTR _________________________________________________________44

5.6.3 Thermoporosimetry vs. NIR ______________________________________________45

5.6.4 Applicability of methods _________________________________________________46

6. Conclusions __________________________________________ 47

References

iv

ACKNOWLEDGEMENTS

This study was carried out at the Pharmaceutical Technology Division, Department of

Pharmacy, University of Helsinki, during the years 1997-2001.

I express my warmest gratitude to Professor Jouko Yliruusi for his support and

encouragement during these years. His enthusiasm for physical and molecular pharmacy

made the completion of this study possible. Furthermore, I want to thank him for letting

me do things in my own way. I am grateful also to my second supervisor, Docent Leena

Hellén, for her warm support, never-ending optimism and positive way of thinking

during all stages of this work.

I am most grateful to Professor Mike Newton, University of London, for giving

me the opportunity to visit his laboratory, which is world-famous in extrusion studies. It

was a privilege to study rheology under his kind guidance. I wish to thank also Dr.

Fridrun Podczeck for the discussions in the field of statistics and data analysis and for

supervising me during the powder rheometer studies.

I am deeply grateful to Ass.Prof. Torben Schæfer, The Royal Danish School of

Pharmacy, for his time and guidance during the mixer torque rheometer studies. His long

experience in the field of granulation combined with his accurate and abundant

comments on the manuscripts taught me a great deal of scientific thinking and writing.

I express my sincerest appreciation to Docent Anne Juppo and Dr. Kirsti

Saarnivaara, the reviewers of this thesis, for their constructive comments and suggestions

for improvement of the manuscript.

Special thanks belong to my dear colleague Dr. Jukka Rantanen for introducing

me the fascinating world of NIR spectroscopy. His never-failing support, encouragement

and friendship have been invaluable during these years.

I express my appreciation to Dr. Thad Maloney, Professor Hannu Paulapuro and

Leena Nolvi (Laboratory of Paper Technology, Helsinki University of Technology) for

their co-operation in the field of thermoporosimetry.

Docent Jukka-Pekka Mannermaa is greatly acknowledged for his support in those

moments when I needed it most.

v

My warm thanks are due to Professor Jouni Hirvonen for revising the language

of the manuscript. Jaakko Herttua is acknowledged for the illustration of the rheometers.

Tero Närvänen is thanked for the specific surface analyses. Lasse Kervinen is

greatly acknowledged for the pleasant discussion during this work. Orion Pharma is

acknowledged for the loan of the FT-NIR instrument.

One important reason why this thesis could be completed was the constant

support of my colleagues at the division. I am most grateful to the whole staff of

Pharmaceutical Technology Division for providing the most pleasant and convenient

environment in which to work (and to party!). I wish to thank especially Karin Krogars,

Amie Kaukonen, Niklas Laitinen and Anna Jørgensen for their friendship and for sharing

the moments of joy and despair with me.

I express my gratitude also to M.Sc. thesis students Flemming Jørgensen and

Krista Mäkelä for their co-operation and friendship.

I am most grateful for the closest friends and relatives here in Finland and abroad

for all the support and understanding during these years.

Finally, my warmest thanks belong to my beloved parents and brother Pekka, for

their unfailing encouragement and loving support.

The financial support from the Finnish Cultural Foundation, the Finnish

Pharmaceutical Society and NorFa are gratefully acknowledged.

Helsinki, May 2001

vi

List of original publications

This thesis is based on the following original papers, which are referred to in the text by

the Roman numerals I-V.

I Luukkonen, P., Schæfer, T., Hellén, L., Juppo, A.M. and Yliruusi, J., 1999.

Rheological characterization of microcrystalline cellulose and silicified

microcrystalline cellulose wet masses using a mixer torque rheometer, Int.J.Pharm.

188 181-192.

II Luukkonen, P., Schæfer, T., Podczeck, F., Newton, J.M., Hellén, L. and Yliruusi,

J., 2001. Characterization of microcrystalline cellulose and silicified

microcrystalline cellulose wet masses using a powder rheometer. Eur.J.Pharm.Sci.

13 143-149.

III Luukkonen, P., Newton, J.M., Podczeck, F. and Yliruusi, J., 2001. Use of a

capillary rheometer to evaluate the rheological properties of microcrystalline

cellulose and silicified microcrystalline cellulose wet masses. Int.J.Pharm. 216 147-

157.

IV Luukkonen, P., Rantanen, J., Mäkelä, K., Räsänen, E., Tenhunen, J. and Yliruusi,

J., 2001. Characterization of silicified microcrystalline cellulose and α-lactose

monohydrate wet masses using near infrared spectroscopy. Pharm.Dev.Technol. 6 1-

9.

V Luukkonen, P., Maloney, T., Rantanen, J., Paulapuro, H. and Yliruusi, J., 2001.

Microcrystalline cellulose-water interaction – a novel approach using

thermoporosimetry. Submitted for publication.

vii

List of abbreviations

BJH Barrett, Joyner, Halenda

CSD colloidal silicon dioxide

DSC differential scanning calorimetry

FBW freezing bound water

FSP fiber saturation point

FT-NIR Fourier transform near infrared

FT-NMR Fourier transform nuclear magnetic resonance

IR infrared

MCC microcrystalline cellulose

MTR mixer torque rheometer

NaCMC sodium carboxymethylcellulose

NFW nonfreezing water

NIR near infrared

NMR nuclear magnetic resonance

PCA principal component analysis

PITs processing-induced transformations

PR powder rheometer

PSD pore size distribution

SMCC silicified microcrystalline cellulose

TBW total bound water

TFW total freezing water

Tg glass transition temperature

TS tensile strength

WRV water retention value

ZTL zero torque limit

1

1. INTRODUCTION

Granulation studies have traditionally focused on the investigations of raw materials and

dry granules. However, it is well known that moisture affects both the physical and

chemical properties of pharmaceutical solids. Depending on the material properties water

can be adsorbed to solid surfaces, absorbed into amorphous solids, condensed into

micropores or it can form crystal hydrates or dissolve the solid (Zografi, 1988). The first

methods to characterise wet mass and to detect the end-point of granulation were torque

and power consumption measurements (Leuenberger et al., 1979; Bier et al., 1979).

However, these methods do not give direct information about the state of the wet mass.

In order to increase the degree of safety and quality of drug manufacturing processes, the

fundamental understanding of processing-induced transformations (PITs) is needed

(Morris et al., 2001). Recently, techniques based on the characterisation of the wet mass

at a molecular level, such as near infrared (NIR) spectroscopy, have been introduced for

in-process control.

Cellulose has a porous structure with both crystalline and amorphous regions.

The interaction of microcrystalline cellulose (MCC) with water is complicated and several

theories have been proposed on how water is bound to cellulose (Froix and Nelson,

1975; Zografi et al., 1984; Fielden et al., 1988; Blair et al., 1990). In the interaction with

water microcrystalline cellulose particles swell, and during drying the cellulose particles

shrink (Kleinebudde, 1994a-b). Microcrystalline cellulose is widely used as an excipient in

drug manufacturing processes, e.g. in direct compression and wet granulation. In

addition, MCC is an essential excipient in extrusion-spheronisation to achieve the

required rheological properties of the wet mass. The properties of microcrystalline

cellulose have been tried to improve by co-processing MCC with colloidal silicon

dioxide. Silicified microcrystalline cellulose (SMCC) should have better compaction

properties and it should be more resistant to wet granulation than conventional MCC

grades (Sherwood and Becker, 1998). Cellulose is essential in paper making process as

well, and the properties of cellulose fibers have been widely studied in the field of pulp

and paper research. In this study some of the theories and techniques used in pulp and

paper research have been applied to the study of MCC and silicified microcrystalline

cellulose (SMCC).

2

In order to understand the behaviour of the wet powder mass in different unit

operations, it is important to understand solid-water interactions and the state of water in

solids. This is especially important if the properties of the powder mass change during

wetting and drying, as is the case with microcrystalline cellulose. Most of the earlier

studies with MCC and SMCC have been made using relatively low moisture contents,

which do not include the moisture contents used in wet granulation and extrusion-

spheronisation. Until now there has been a limited number of methods to study the

properties of the wet masses and the methods have not been compared.

3

2. THEORY

2.1 Water-solid interactions

2.1.1 Structure of water

The water molecule consists of an atom of oxygen, which is covalently bound to two

atoms of hydrogen. In addition to these covalent bonds the oxygen atom has two lone

electron pairs. It has been found that a specific attraction exists between the

electronegative atoms (oxygen) and hydrogen atoms, particularly when the latter are

themselves chemically bonded directly to the electronegative atoms. The fact that water

can act both as a donor and as an acceptor of hydrogen atoms creates a great potential

for hydrogen bonding. The manner of hydrogen bonding in water is co-operative (Frank

and Wen, 1957). This means that the presence of one hydrogen bond facilitates the

formation of additional bonds.

It was earlier believed that there are no hydrogen bonds between water molecules

in the water vapour (Choppin and Downey, 1972). However, Dyke et al. (1977) showed

that dimers (H2O)2 exist in the vapour phase with a linear structure in the hydrogen bond.

Recently even water clusters have been found in water vapour (Jeffrey, 1997). The length

of the hydrogen bond (~2.95 Å) in the water vapour is longer than the observed

distances in liquid water (~2.85 Å) and in ice (~2.74 Å), which indicates that the

hydrogen bonds are weaker in water vapour.

2.1.2 Water sorption by solids

The physical and chemical properties of pharmaceutical solids are dependent on the

presence of moisture. Residual water may have significant effects on glass transition

temperature, stability, powder flow, compaction, and dissolution rate (Oksanen and

Zografi, 1990). The residual water may exist due to a prolonged exposure to water

vapour or as a result of processing that involves the use of water. The effects of moisture

depend on the amount of water sorbed and desorbed.

In order to understand the effects water can have on the properties of solids, the

state of water in solids, and the different mechanisms of water sorption in different

materials should be understood. Interactions include adsorption to solid surfaces,

absorption into amorphous solids, formation of crystal hydrates, condensation into

4

micropores and dissolution of solid in sorbed water (Zografi, 1988). The mechanism of

water sorption is greatly affected by the crystallinity of the solid material. In addition to

crystallinity, water-solubility, porous structure and the ability to form crystal hydrates

determine the mechanism of water sorption into the solid.

The most common interaction of water with surfaces is hydrogen bonding

through the oxygen atom (Thiel and Madey, 1987). Water molecules initially adsorbed on

the surfaces usually form a monomolecular layer. However, hydrogen bonded water

clusters may form even before the monomolecular layer is completed (Thiel and Madey,

1987; Zografi, 1988). This is due to the hydrogen bonding between two or more water

molecules, which is often energetically competitive with the molecule-substrate bond.

Furthermore, Pizzi et al. (1987a,b) and Berthold (1996) proposed that different polar

groups determine the magnitude of water an adsorption site would attract. According to

this theory, certain adsorption sites can adsorb one, two and, in some cases, more water

molecules before the other sites adsorb their first water molecule.

The adsorption of water to nonhydrating crystalline solids depends on the

polarity of the surfaces and is proportional to the surface area (Kontny and Zografi,

1995). The amount of moisture sorbed by amorphous solids is typically much greater

than that sorbed by crystalline substances. Similarly to crystalline solids, water is first

adsorbed to the surfaces of material. As more water molecules adhere to the surface,

moisture transfers into the amorphous material (Young and Nelson, 1967a,b). The

greater the chemical affinity of water for the solid, the greater the total amount that can

be absorbed. As soon as water penetrates into the amorphous solid structure, it acts as a

plasticizer and reduces the glass transition temperature, Tg (Froix and Goedde, 1976;

Zografi, 1988). Water replaces crosslinking hydrogen bonds between the neighbouring

polar groups of the solid, and thus loosens the structure of the solid.

Solids that form crystal hydrates adsorb initially small amounts of water to their

external surface. Crystal hydrates may be formed either in amorphous or crystalline

solids. In the case of a crystalline solid water molecules are able to penetrate into the

crystal lattice in a well-defined molecular position within the unit cell, and to form

hydrogen bonds to certain groups with a specific stoichiometry (Byrn et al., 1999). The

strength of the water-solid interaction depends on the level of hydrogen bonding

possible within the lattice (Zografi, 1988; Jeffrey, 1997; Byrn et al., 1999). In some cases

5

(e.g. caffeine and theophylline), where hydrogen bonding is relatively weak, water

molecules form hydrates mainly due to their space-filling role.

Water sorption onto porous solids differs from water uptake onto the surfaces of

flat materials. This is generally attributed to the increased attractive forces between

adsorbate molecules that occur as surfaces become highly curved such as in a pore or

capillary (Kontny and Zografi, 1995). At first, water is adsorbed onto the surface of the

pore wall, after which it is condensed and fills the core of the pore. This phenomenon is

referred to as capillary condensation and is described by the Kelvin equation [Eq. 1]

presented originally by Thomson (1871):

rRTV

PP mr γ2ln0

−= [1]

where rP is the vapour pressure in a pore radius r, 0P is the vapour pressure on flat

surfaces, γ is the surface tension of the adsorbed film, mV is the molar volume of the

liquid, R is the gas constant, and T is the temperature.

Crystalline, non-hydrate forming, water-soluble solids have a tendency to

deliquesce, i.e. to dissolve in their own sorbed water. At low moisture contents water is

adsorbed to the surface of the solid. As moisture content reaches the critical relative

humidity, the solid begins to dissolve in the sorbed film of water (Van Campen et al.,

1983; Kontny and Zografi, 1995). Deliquescence arises because of the high water

solubility of the solid and the significant effect it has on the colligative properties of

water. The process of water uptake will continue until the entire solid has dissolved.

2.1.3 Wet agglomeration

The mechanisms of bonding in the wet state depend on capillary and interfacial forces

between the particles. Once sufficient amount of liquid is added, the granulation shifts

from an immobile surface liquid state to a mobile liquid film state. Newitt and Conway-

Jones (1958) defined this theory of granulation in three different states (pendular,

funicular and capillary), into which Barlow (1968) added the fourth state (droplet) (Fig.

1a-d). Each of the four states represents a progressive increase in the moisture content,

with a corresponding change in capillary forces until the droplet state is reached.

At low moisture content water is held in the granule as discrete lens-shaped rings

at the points of contact of the particles (Fig. 1a). Particles are held together by surface

6

tension at the solid-liquid-air interface and the hydrostatic suction pressure of the liquid

bridge. At somewhat higher moisture content the rings coalesce and there is a continuous

network of liquid interspersed with air (Fig. 1b). With further increase in the water

content all the pore spaces in the granule are completely filled, and concave menisci

develop at the surface of the agglomerate (Fig. 1c). The cohesiveness of the wet mass

increases with increasing moisture content. The droplet state (Fig. 1d) occurs when the

liquid completely surrounds the granule, resulting in an external liquid phase, with an

internal solid phase (Barlow, 1968). In the droplet state, only surface tension holds the

drop together and there are no longer any internal interfacial forces.

Figure 1. Liquid saturation model. a) pendular b) funicular c) capillary and d)droplet state according to Newitt and Conway-Jones (1958) and Barlow (1968).

The tensile strength in the pendular state is about one-third of that in the

capillary state, and the tensile strength of agglomerate in the funicular state takes an

intermediate value between that of the capillary and pendular states (Rumpf, 1958). The

tensile strength of agglomerates may be estimated by the following equation [Eq. 2]:

θγε

ε cos1d

SCTS −= [2]

where TS is the mean tensile strength, S is the liquid saturation, C is a constant

depending on the particle shape, ε is the porosity of the agglomerate, γ is the surface

tension of the binder liquid, d is the diameter of particles, and θ is the contact angle of

the binder liquid to the solid.

The liquid saturation depends on the amount of liquid and the intragranular

porosity according to the equation [Eq. 3] (Kristensen et al., 1984):

( ) ρε

ε−= 1HS [3]

a) b) c) d)

7

in which H is the ratio of the mass of liquid to the mass of solid particles, ε is the

intragranular porosity and ρ is the particle density. The equation presumes that the

liquid density is unity.

Intragranular porosity is defined according to equation [Eq. 4]:

s

b

ρρε −= 1 [4]

where bρ is the apparent or bulk density of the packing and sρ is the true density of

solid particle.

According to Sastry and Fuerstenau (1973) the granule growth mechanisms

include the nucleation of fine powder to form initial primary granules, the coalescence of

existing granules, the layering of raw material onto the previously formed nuclei or

granules and the material transfer by abrasion. Many researchers have observed that the

rate of granule growth is strongly dependent upon the liquid content of the granulating

mass (Newitt and Conway-Jones, 1958; Kristensen and Schæfer, 1987). This is attributed

either to the increased granule plasticity or to the surface moisture leading to a greater

probability that the granules will stick together in collision. When one or more of the

components are soluble in water, the total volume of the liquid or solution phase rather

than the moisture content controls the granulation behaviour (Sherrington, 1968).

In pharmacy the wet granulation is used primarily for the preparation of materials

for tabletting. The amount of liquid required for an uncritical granulation process

depends on a large number of factors, such as material properties, liquid characteristics

and the mode of action of the equipment (Capes, 1980; Kristensen and Schæfer, 1987).

Since the consolidation of moist agglomerates is a basic mechanism involved in the

granule growth, equipment with high shear forces require less water than equipment with

low shear forces (Kristensen and Schæfer, 1987). Newitt and Conway-Jones (1958) noted

that the critical moisture content required for granulation correlated with 90% of the

moisture required to saturate the voidage in the powder mass.

Extrusion-spheronisation process is a special case of wet granulation, which

requires more water as compared to traditional granulation. The specific requirements for

wet masses will be discussed in chapter 2.2.5. The amount of water required for

successful extrusion-spheronisation has been traditionally determined by trial and error.

However, the level of water at the capillary state should be comparable with that found

8

for the optimum production of pellets during spheronisation (Miyake et al., 1973; Rowe

and Sadeghnejad, 1987).

2.2 Special characteristics of cellulose with water

2.2.1 Structure of cellulose

Two major sources of cellulose fibers are cotton seed hair and wood. Cotton is almost

pure cellulose, while wood contains cellulose (40-50%), hemicelluloses (20-30%) and

lignin (20-30%). The role of cellulose is to reinforce the wood fiber. Lignin adds strength

to the structure and hemicelluloses bind the cellulose and lignin together. In the pulping

process the cellulose fibers are separated and part of the lignin and the hemicelluloses are

dissolved. ”Dissolving pulp” is the trade name for pure cellulose fibers, which are used in

chemical processes.

The cellulose polymers are built up of cellobiose units, which are linked together

by ß-(1-4) glucosidic bonds (Fig. 2). The cellobiose units (A) form larger unit cells (B),

which form crystallites (C). Crystallites with crystalline and amorphous regions are

combined into microfibrils (D) and fibrils (E), which finally form cellulose fibers (not

shown).

Figure 2. The structure of cellulose.

Microcrystalline cellulose is manufactured from dissolving pulp by acid

hydrolysis. After washing with water the suspension is spray dried into microcrystalline

powder. The manufacturing process shortens the cellulose chains and removes the

dissolving parts of cellulose fibers, i.e. hemicelluloses and lignin. During the process the

cellulose fibers (length 1-3 mm) are reduced to smaller particles, which diameter

(normally 50-100 µm) can be defined in the spray drying process.

HO

O

OH

OH

O

HO

OH

O

AA B C D E

9

Milling of dissolving pulp produces powdered cellulose. The degree of

polymerisation is much greater in powdered cellulose (~500) than in microcrystalline

cellulose (~220). In addition, the degree of crystallinity of powdered cellulose (15-45%) is

smaller compared to that of microcrystalline cellulose (65-75%). This means that the

cellulose chains are much longer in powdered cellulose and the structure of powdered

cellulose contains more amorphous regions than that of MCC.

Silicified microcrystalline cellulose (SMCC) is produced by co-processing

colloidal silicon dioxide (CSD) together with MCC, i.e. CSD is added to the MCC slurry

before spray drying.

Colloidal grades of MCC are produced by co-processing MCC with the

hydrophilic sodium carboxymethylcellulose (NaCMC). Colloidal grades of MCC vary in

terms of the amount of NaCMC added, the viscosity of NaCMC and the drying method

used during manufacturing.

2.2.2 Water sorption by cellulose

Hydrogen bonds within and between cellulose chains give the cellulose molecule a stiff

and rigid nature. In the initial stage of cellulose-water interaction, water molecules are

adsorbed on the cellulose surfaces and bound to the free hydroxylgroups of cellulose

chains by hydrogen bonds. The heat of adsorption provides a source of energy for the

breaking of lateral bonds between the polysaccharide molecules (Emerton, 1980).

Consequently, water is able to replace the crosslinking hydrogen bonds between cellulose

chains, and thus to loosen the structure of cellulose (Fig. 3). At high relative humidities,

capillary condensation occurs and the porous cellulose fibers can hold appreciable

amounts of water in submicroscopic pores, cracks and crevices between fibrils. To some

extent water is held also in pores and gaps between the fibers (Weise, 1997). In addition

to the water adsorption to the surfaces, water is also absorbed into the amorphous parts

of cellulose.

Water in wood pulps has been categorised according to the thermodynamic

properties to nonfreezing, freezing bound and free water (Nakamura et al., 1981;

Hatakeyama et al., 1983; Maloney, 2000). Nonfreezing water is a type of bound water,

which does not show any first order transition down to –170ºC. Nonfreezing water

consist of the surface water, which is hydrogen bonded to the cellulose, and the further

layers of water molecules, which are still significantly influenced by the cellulose surface

10

(Li, 1991). Freezing bound water is located in submicroscopic pores, cracks and crevices

between fibrils due to the capillary condensation. Water with thermodynamic properties

similar to those of pure water is called free water. Water can be both inside and outside

fibers, since the methods based on thermodynamic properties cannot differentiate the

location of water.

There is still not a well-accepted thermodynamic theory to explain why some of

the water in hydrated systems does not freeze. However, it has been shown that there is a

correlation between the number and type of available hydration sites and the quantity of

nonfreezing water (Nakamura et al., 1981; Berthold, 1996). NMR measurements have

shown that vicinal water (water near a surface) is in a high state of motion and in this

respect, the term bound may be misleading (Li, 1991).

Figure 3. a) Hydrogen bonding between two adjacent cellulose molecules, b)Hydrogen bonding of two cellulose molecules through a monolayer of watermolecules and c) Irregular hydrogen bonding of two cellulose molecules,resulting from sorption of multilayers of water molecules (Emerton, 1980).

b)

a)

c)

11

2.2.3 Swelling

In the interaction with water new pores are generated as the cellulose particles swell. In

the pulp fiber literature, swelling is usually defined as the amount of water contained in

the saturated fiber cell wall. Liquids that are able to penetrate into the crystalline regions

induce intracrystalline swelling of cellulose (Zeronian, 1985). Since water does not

penetrate well-defined crystalline zones, it can cause only intercrystalline swelling of

cellulose. The most popular techniques to measure the swelling of pulp fibers have been

water retention value (WRV) and solute exclusion technique (Stone et al., 1968; Scallan

and Carles, 1972; Weise, 1997; Maloney, 2000). WRV determines the moisture content of

pulp fiber after centrifugation. Solute exclusion technique measures the fiber saturation

point (FSP), i.e. the total amount of water inside the cellulose particles. High molecular

weight dextran solution, which cannot penetrate cellulose particles, is added to the wet

cellulose mass. Water outside the cellulose particles is able to dilute the dextran solution,

and the method is based on the measurement of the concentration change of dextran

solution.

In the paper making process swelling of pulp fibers can be increased by beating,

which cuts cellulose fibers and causes morphological changes. Consequently, the lamellar

structure of fibrils inside the cellulose fibers is loosened, which increases the water

penetration into the cellulose structure (Stone et al., 1968; Emerton, 1980; Li, 1991;

Maloney et al., 1998).

2.2.4 Hornification

Most of the pores, which have been formed in the interaction with water, will collapse

during dewatering of cellulose. Urquhart (1924) suggested that the number of polar

hydroxyl groups of cellulose fibers is decreased during rewetting because some of them

have formed hydrogen bonds between themselves. Jayme (1944) named this

phenomenon as hornification, and defined it as a decrease in water retention value. The

closure of capillaries progresses as if closing a zipper, which might not be reopened by

newly intruding water. Despite its original definition, hornification is commonly used as a

descriptive term for the physical and chemical changes that occur in pulp fibers during

drying, principally shrinkage and formation of internal hydrogen bonds (Weise, 1998).

The phenomenon is seen also in pharmaceutical wet granulation as a remarkable increase

in the density of the granules (Chatrath, 1992; Sherwood and Becker, 1998; Habib et al.,

1999).

12

2.2.5 Behaviour of MCC/SMCC in extrusion-spheronisation

MCC is an essential excipient in extrusion-spheronisation to achieve the required

rheological characteristics of the wet mass. The wet mass must be brittle enough to cut

into small pieces in the spheronizer but on the other hand it should be sufficiently plastic

to enable the moulding of the extrudate into a sphere (Fielden and Newton, 1992).

Fielden et al. (1988) and Ek and Newton (1998) suggested that microcrystalline cellulose

can be described as a molecular sponge, since MCC is able to trap a large volume of

water within its porous structure and retain it despite the application of high pressures.

During extrusion the sponges are compressed until water is squeezed out and lubricate

the particle flow through the extruder.

Kleinebudde (1997) proposed that MCC particles are broken down into smaller

particles by shear forces acting on the particles during granulation and extrusion. The

new model was named as the crystallite-gel model. According to this model finally single

crystallites of colloidal size prevail, and these single particles, in the presence of water, are

able to form a gel and immobilise water. Recently, Kleinebudde et al. (2000) postulated

that the sponge model is more appropriate for the cellulose type a with high degree of

polymerisation (powdered cellulose), whereas the gel model is more applicable to

cellulose types with a lower degree of polymerisation (MCC and SMCC).

Extrusion-spheronisation consists of three different stages; i.e. granulation,

extrusion and spheronisation. Every step in the process of extrusion-spheronisation

affects the properties of the resulting pellets (Schmidt and Kleinebudde, 1999). However,

only few authors have reported studies concerning the impact of the granulation step.

When comparing planetary and high-shear mixers, Baert et al. (1991) and Vervaet (1997)

found out that granulation in the high-shear mixer and increased granulation times

resulted in higher extrusion forces for the same composition. Furthermore, Schmidt and

Kleinebudde (1999) found out that higher shear forces during granulation resulted in a

higher water content necessary for a successful pelletisation. Since the results were in

contrast to conventional granulation, they explained their results with the crystallite-gel-

model. According to the model, high shear forces during granulation and extrusion

should result in a more delicate network, which requires more water to obtain the same

plasticity as compared to a coarser network of the gel.

13

2.3 Wet mass characterisation

2.3.1 Torque methods

The first attempts to follow the granulation process were made by measuring

either the power consumption or torque of a planetary mixer (Lindberg et al., 1974;

Travers et al., 1975). A few years later Leuenberger and his colleagues compared both

methods by measuring the torque and electrical power consumption simultaneously

(Leuenberger et al., 1979; Bier et al., 1979). They found out that both the variables

changed as a result of a change in the cohesive force or in the tensile strength of the wet

agglomerates. The authors were able to connect the torque curve to the liquid saturation

and to the increased particle size of the wet powder mass, i.e. they were able to use the

torque curve to the end-point detection of the granulation process (Fig. 4). According to

Leuenberger et al. (1981) and Imanidis (1986), granules for tabletting can be produced in

the third phase (S3-S4). Recently, the torque measurements have been used to control the

end-point of a wet pelletisation process in a rotary processor by Kristensen et al.

(2000a,b).

Figure 4. Torque curve connected to the liquid saturation (Imanidis, 1986).

A π-value derived from a power consumption/torque curve [Eq. 5] has been

used to normalise the uncritical amount of liquid for granulation (Imanidis, 1986; Usteri

and Leuenberger, 1989) and for extrusion-spheronisation (Nürnberg and Wunderlich,

1999a-b).

( ) ( )252 SSSS −−=π [5]

0

100

200

300

400

0 5 10 15 20 min

Phase VPhase IVPhase III

Phase II

S2 S3 S4 S5

Torque

14

where S is the amount of granulating liquid, 2S is the amount of granulating liquid

necessary (which corresponds to a moisture equilibrium at approx. 100% relative

humidity) and 5S is the complete saturation of interparticulate void space before a slurry

is formed.

In addition to the instrumentation of different granulators, various equipment

have been introduced for the evaluation of wet massing behaviour of materials. Alleva

(1984) used a granulation rheology apparatus and Staniforth et al. (1988) a modified shear

cell to study the cohesiveness of wet MCC masses. Malamataris and Kiortsis (1997)

introduced a wet-kneading rheometer to study the funicular phase of liquid saturation.

The mixer torque rheometer (MTR) has been used to study the source variation of MCC

grades (Rowe and Sadeghnejad, 1987; Parker and Rowe, 1991), binder-substrate

interactions (Parker et al., 1990; Parker et al., 1991), and the mixing kinetics of water and

MCC (Hancock et al., 1992). Recently the torque measurements have been used for

scale-up and process control (Landín et al., 1996; Faure et al., 1999) and for the

evaluation of the optimum water level for extrusion-spheronisation (Souto et al., 1998;

Chatlapalli and Rohera, 1998; Luukkonen et al., 1999).

The mixer torque rheometer has two intermeshing

paddles, and the sample is compressed and expanded

between contrarotating blades with a changing gap (Fig. 5).

The reaction of the mixing bowl is continually recorded via

a torque arm fixed to the main body of the mixer and

linked to a calibrated dynamometer. The MTR measures

two different torque parameters, the amplitude of the

oscillations (torque range) and the mean torque increase from the baseline (mean torque).

The mean torque describes the mean resistance of the mass to mixing and the torque

range reflects the rheological heterogeneity of the mass. In the MTR experiments the

material exhibits an increase in torque with increasing water content as the consistency of

the wet mass rises to a maximum, whereafter it decreases as the material becomes

overwetted. Alleva (1984) and Rowe and Sadeghnejad (1987) have suggested that this

behaviour is consistent with the different states of liquid saturation defined by Newitt

and Conway-Jones (1958).

The newest equipment to measure the wet mass consistency is a powder

rheometer, which should be able to measure both the dry and wet powder masses. The

Figure 5. Mixertorque rheometer.

15

equipment consists of a cylindrical measuring vessel (50 x 140 mm) and a rotor blade

which descends into the cylinder in a helical path (Fig. 6). The helical path is defined as

the spiral movement of the impeller blade tips during downward or upward movement

of the rotating impeller. By changing the impeller direction of rotation and the helical

path, the impeller blade causes either compaction or slicing of the powder bed. The

equipment enables the user to choose a combination of

rotor path settings and blade tip velocities. The apparatus

allows the measurement both of the force exerted by the

powder on the blade and the resistance of the powder to

movement (torque). Each property is recorded as a

function of the height of the blade in the powder bed.

Powder rheometer has been used so far only to

study the capsule filling properties of MCC and

granulated powdered cellulose (Podczeck, 1999a-b;

Podczeck and Newton, 2000). The authors found out

that the powder rheometer was able to identify

similarities and differences in the flow, shear and packing

properties of dry powders.

2.3.2 Rheological methods

The rheological properties of an extrudate depend on the properties of the material and

the liquid content of the wet mass (Fielden and Newton, 1992). If the wet mass is too

dry, the plasticity of the extrudate is insufficient and the extrudate does not round in the

spheronisation. On the other hand, if the wet mass is too wet the pellets will agglomerate

during spheronisation. To design an optimal formulation for extrusion-spheronisation, it

is necessary to characterise the mechanical and rheological properties of the wet mass (Li

et al., 2000). Since the wet mass undergoing extrusion and spheronisation experiences

different kinds of stresses, it is apparent that the rheological or mechanical properties of

the wet mass may address the requirements of both the processes (Shah et al., 1995).

However, there is a lack of understanding of these processes and of the requisite

properties of materials and formulations. Only few studies have considered the

rheological properties of wet masses that make them suitable for extrusion-

spheronisation.

Figure 6.Powder rheometer.

16

According to Ovenston and Benbow (1968), the parameters necessary to

characterise the flow of a material through a die during extrusion can be measured

directly only using a capillary rheometer. The ease of extrusion depends on the excess of

the liquid present over that what is required to fill the interparticle space between the

particles (Benbow and Bridgwater, 1993). Powders which have a broad particle size

distribution, pack closely requiring less liquid to fill the spaces between particles than

powders having a narrow particle size distribution (Benbow et al., 1987; Chen et al.,

2000). In addition to particle size and size distribution, particle shape and the state of

surface exert significant influence on the extrusion behaviour of pastes (Chen et al.,

2000).

The extrusion of pastes can result in the phenomenon of liquid migration, which

causes changes in fluid content within the wet mass. Water movement in the material

and its liquid retention potential has been studied using pressure membrane technique

(Fielden et al., 1992; Knight, 1993; Boutell, 1995) and centrifugation (Tomer and

Newton, 1999a; Tomer et al., 2001). Water movement during extrusion has been

followed by measuring the moisture content of extrudate fractions (Harrison, 1982; Baert

et al., 1992; Knight, 1993; Tomer and Newton, 1999b). Moisture distribution of

extrudate plugs can be determined also by magnetic resonance imaging (MRI) (Tomer et

al., 1999a;b).

Ram extruder and compresso-rheometer have been used to screen the materials

for successful extrusion-spheronisation (Harrison et al., 1985; Delalonde et al., 1997;

Tuleu and Chaumeil, 1998; Delalonde et al.,

2000). When a ram extruder has been used as

a capillary rheometer, information can also be

obtained about the elastic behaviour of wet

masses (Chohan and Newton, 1996). Shah et

al. (1995) combined rheograms (yield values),

tensile strengths and yield loci with data from

extrusion to define a rheological window

within which both the extrusion and

spheronisation could be successfully carried

out. Li et al. (2000) introduced a triaxial

compression test for pharmaceutical wet

Load cell

Piston

Barrel

Wet mass

Die

Extrudate

Figure 7. Ram extruder.

17

masses. They were able to determine stress-strain curves, pore water pressure, shear

strength and cohesion together with an angle of internal friction for the wet powder

masses.

The ram extruder consists of a barrel, a die and a piston (Fig. 7). Wet mass is

packed into the barrel and compressed by hand pressure, after which the ram extruder is

driven downwards by a mechanical press. The air is expelled during the compression

stage, and in the beginning of extrusion the system consists of liquid and solid particles

only. The resistance to movement involves two components; the convergent flow from

the wide barrel to the narrow die and the flow through the die.

When different ram velocities and dies with different lengths are used, ram

extruder works as a capillary rheometer. The die wall shear stresses can be derived from

the plots proposed by Bagley (1957) according to equation [Eq. 6]:

LPRw 2∆=τ [6]

in which wτ is wall shear stress, L is the die length, R is the die radius and P∆ is the

pressure difference between the ends of the die.

The pressure drop along the barrel can be estimated by determining the values of

piston pressure against the length to radius ratio of the die, and extrapolating to a zero

value. A finite pressure loss at the zero die length can thus be estimated and this is

equivalent to the drop of pressure along the barrel. Thus, the equation [6] can be

modified as:

( ) LRPPw 20−=τ [7]

where P is the piston pressure and 0P is the upstream pressure loss.

The die wall shear rates are calculated according to equation [Eq. 8]:

3

4RQ

A πγ = [8]

where Q is the volumetric flow rate and R is the radius of the die.

The die wall shear stresses can then be plotted against the die wall shear rates to

construct a flow curve, which provides a representation of the rheological properties of a

given material.

18

2.3.3 Thermal methods

Differential scanning calorimetry (DSC) has been traditionally used to measure the

nonfreezing water in starch, cellulose and pulp fibers (Mousseri et al., 1974, Nelson,

1977; Nakamura et al., 1981; Hatakeyama et al., 1983). The amount of nonfreezing water

is calculated by subtracting the total freezing water in the sample (determined from the

measured heat) from the total water in the sample (determined gravimetrically). The

freezing bound water is much more difficult to quantify than the nonfreezing water,

because the DSC melting peak overlaps with that of the bulk water (Nakamura et al.,

1981; Yamauchi and Murakami, 1991; Maloney, 2000). However, the water held in the

capillaries of porous materials has a depressed melting point. This is due to the higher

pressure of water in the cavities with a curved interface. The depression of melting

temperature of a material has been used to measure the pore size distribution by

thermoporosimetry. Generally, thermoporosimetry has been applied to systems where

the pores are sufficiently small, so that separate bulk and pore water peaks are formed in

the melting endotherms (Ishikiriyama and Todoki, 1995). For water saturated pulp fibers,

the melting temperature of the water in the cell wall is depressed only slightly below that

of the bulk water, so the pore and bulk water peaks overlap making the analysis difficult.

In order to measure the

pore size distribution of pulp

fibers with DSC, Maloney et al.

(1998) developed an isothermal

step melting procedure. The

principle of the isothermal melting

technique is to raise the tempera-

ture in a frozen sample to a

present value, where it is held

constant until the melting

transition is completed (Fig. 8).

The melting of the sample is then

repeated at a slightly higher

temperature. The heat absorbed at

each temperature is measured by

Time

Tem

pera

ture

Conventional DSC ProgramStep Program

t∆

isott ∆∆=β

isot∆

Figure 8. Comparison of dynamic andisothermal step program. β is the rate of thedynamic segment, T∆ is the temperatureincrement of the steps and isoT∆ is theduration of the isothermal segment (Maloney,1999).

19

integrating the endotherm. The melting heat is assumed to correlate directly with the

amount of melted water and, consequently, the pore size distribution is calculated from

the melting temperature depression of the imbibed water.

The relationship between pore diameter (D) and the melting temperature

depression is described by the Gibbs-Thomson equation [Eq. 9]

TH

TvDf

lsm

∆−=

γ04 [9]

where mv is the molar volume of ice, 0T is the melting temperature of water at a normal

pressure, lsγ is the surface energy at the ice-water interface (12.1 mJ/m2), fH is the

specific heat of melting of bulk water (334 J g-1) and T∆ is the melting temperature

depression (°C).

2.3.4 Molecular methods

Bugay and Williams (1995) stated that one stage of the characterisation of pharmaceutical

solids must be at a molecular level. Suitable methods for this purpose are the various

forms of molecular spectroscopy techniques such as infrared, Raman, and nuclear

magnetic resonance (NMR). Infrared and Raman analyses provide a complete vibrational

motion analyses of the molecule, whereas NMR provides insight into the local

environment of each NMR active atom.

The basic idea of the NMR method is that the molecular dynamics of liquid

molecules in the vicinity of a solid surface is perturbed and, therefore, the relaxation

times for the liquid molecules near the surface are different from that of the bulk liquid.

NMR has been used to study the state of water in cellulose samples (Froix and Nelson,

1975; Froix and Goedde, 1976; Peemoeller and Sharp, 1985), the pore structure of

porous cellulose beads (Ek, 1995) and the pore size distribution of pulp cellulose fibers

(Li, 1991; Maloney et al., 1997).

In Raman spectroscopy, the sample is irradiated with monochromatic laser

radiation, and the inelastic scattering of the source energy is used to obtain a vibrational

spectrum of the analyte. A vibrational mode is infrared-active when there is a change in

the molecular dipole moment during the vibration, whereas the vibrational mode is

Raman-active when there is a change in polarisability during the vibration. The Raman

spectrum of water is one broad, weak band at 3500 cm-1, which does not disturb Raman

20

measurements. Due to this Raman spectroscopy has not been used to study the state of

water in wet masses, but rather the changes occurring in the molecular structure of a

material. These studies include pseudopolymorphic transitions during storage (Taylor and

Langkilde, 2000) and wet granulation (Jørgensen et al., 2001).

Water absorbs mid-IR light strongly and, thus, aqueous samples may be

measured only as thin films. Absorption bands in the near infrared (NIR) region arise

from overtones or combinations of overtones originating in the fundamental mid-IR

region. Overtone bands in the NIR region are typically 10-1000 times less intense than

their corresponding fundamental absorption bands. On the contrary to mid-IR

spectroscopy, near infrared spectra may be obtained noninvasively from samples inside

glass containers or remotely using fiber-optics. In addition, NIR spectroscopy is

particularly well suited to the study of water in carbohydrates, because the specificity of

NIR stretching vibrations involving -CH or -OH groups is high.

The near-infrared spectroscopy absorption spectrum of pure water consists of

five bands with the maxima at 760, 970, 1190, 1450 and 1940 nm (Curcio and Petty,

1951). The OH stretching frequency in the IR region represents the strength of the

hydrogen bonds. Thus, the frequencies and intensities of water bands alter also in the

NIR region with the changes in the strength of hydrogen bonds and hydration (Fornés

and Chaussidon, 1978; Iwamoto et al., 1987; Osborne et al., 1993). Non-hydrogen-

bonded -OH groups have higher stretch frequencies and, hence, a lower wavelength than

the hydrogen-bonded -OH groups. The most intense absorption bands of water are

found around 1450 and 1940 nm (Curcio and Petty, 1951; Osborne et al., 1993). Since

the first overtones of carbohydrates (i.e. cellulose and lactose) and the -OH groups of

water overlap on the same wavelengths (around 1450 nm), the combination band

(around 1940 nm) provides a better selectivity for water and, hence, is the most

important absorption band in the NIR region (Osborne et al., 1993).

NIR spectroscopy has been used previously in the study of the different states of

water in foods (Iwamoto et al., 1987; Osborne et al., 1993) and in cellulose (Delwiche et

al., 1991; Berthold et al., 1998; Buckton et al., 1999; Svedas, 2000). Recently, NIR

spectroscopy has been used to study the pseudopolymorphic transformations (Buckton

et al., 1998; Lane and Buckton, 2000; Räsänen et al., 2001). In wet granulation processes

the NIR methods have been used for on-line moisture detection and end-point

determination (Watano et al., 1990; White, 1994; Frake et al., 1997; Rantanen et al., 1998,

21

Rantanen et al., 2000a,b; Morris et al., 2000). In addition, an effort has been made to use

NIR and water absorbing potential of a material to predict the suitable amount of water

for wet granulation (Watano et al., 1996; Miwa et al., 2000).

NIR spectroscopy is particularly suitable for multivariate data analysis, and

together they form a powerful tool for analysing various materials with widely divergent

properties (Antti, 1999). By using principal component analysis (PCA), the number of

variables is reduced from over a thousand wavelengths to a few principal components

describing the variation in the spectra. Berthold et al. (1998) used NIR spectroscopy in

combination with PCA to study the interactions between water molecules and

carboxymethylated cellulose. The same combination has been used to characterise

different pulpwood samples (Antti, 1999).

22

3. AIMS OF THE STUDY

The aim of the present study was to investigate the rheological properties and the state of

water of microcrystalline cellulose (MCC) and silicified microcrystalline cellulose (SMCC)

wet masses. Furthermore, the purpose was to understand the solid-water interactions and

to compare the characterisation methods available for wet powder masses. The specific

aims were:

1. to investigate the rheological properties of SMCC wet masses in comparison to

the standard grades of MCC using a mixer torque rheometer (MTR) and to

correlate the results with the physical properties of the materials

2. to evaluate the powder rheometer in studying MCC and SMCC wet powder

masses and to make comparisons with the MTR

3. to study the influence of MCC type and water content on the rheological

properties of the wet powder masses using capillary rheometer and to compare

the results with the results obtained by MTR

4. to investigate the energetic state of water in different materials using near infrared

(NIR) spectroscopy and to evaluate NIR spectroscopy in comparison with MTR

5. to study the state of water and the effect of wet massing on different water

fractions of MCC/SMCC wet masses using thermoporosimetry together with

solute exclusion technique and to make comparisons with NIR spectroscopy.

23

4. EXPERIMENTAL

4.1 Materials

Three different grades of microcrystalline cellulose (MCC) were studied: Avicel PH 101

manufactured by FMC Co. (Cork, Ireland) and Emcocel 50 and Prosolv 50 from

Penwest Pharmaceuticals Co. (Nastola, Finland). Prosolv 50 is silicified microcrystalline

cellulose (SMCC) with a 2% w/w silicon dioxide concentration.

In addition, α-lactose monohydrate (Pharmatose 200 M, DMV, Veghel, The

Netherlands), 1:1 mixture of SMCC and α-lactose monohydrate and glass ballotini (40

µm, Jencons Ltd., England) were used in NIR and MTR measurements (IV).

4.2 Characterisation of cellulose powders (I)

The particle size and size distribution were studied using a laser light diffractometer

(Malvern 2600c, Malvern Instruments Ltd., UK) equipped with dry powder feeder.

The specific surface area of each sample was determined in a Coulter SA 3100

apparatus (Coulter, Miami, FL). The samples were first degassed under vacuum at 40ºC

for 24 h. The pore volume size distributions were calculated from the nitrogen

adsorption isotherms by the BJH model (Barrett et al., 1951).

The flowability was determined by an automatic flowability recorder (constructed

at Orion Corporation, Finland, i.e., a non-commercial equipment). The angle of repose

was determined by measuring the angle of the powder pile.

The poured and tapped densities of the materials were determined according to

the test of apparent volume (Ph.Eur. 2nd Ed.). True density measurements were made

using a helium pycnometer (type Accupyc 1330, Micromeritics, Dunstable, UK).

The swelling volume was determined using a method described by Podczeck and

Révész (1993). 500 mg ± 0.5 mg of each material was placed in a 20 ml graduated test

tube with 10.0 ml of distilled water. The solids were suspended by vigorous shaking and

allowed to settle down until the volume was constant. The swelling volume was noted

and the relative increase in swelling was calculated on the basis of the poured bulk

volume.

All the tests mentioned above were made in triplicate.

24

Contact angles were determined using Optical Contact Angle Meter (CAM 200,

KSV Instruments Ltd, Finland). The results present the mean of 20 replicates.

4.3 Preparation of wet masses

The materials were granulated in a planetary mixer (Kenwood Chef Classic, KM 400, UK

(II, V); Hobart, London, UK (II, III); Stephan UMC 5 Electronic, Stephan, Germany

(IV)). Glass ballotini (IV) and ungranulated cellulose wet masses (V) were prepared by

adding the water to the dry powder in a mortar. The wet masses were allowed to

equilibrate overnight in plastic bags before experiments.

4.4 Mixer torque rheometry (I, II, IV)

The rheological profiles of the wet masses were monitored using a mixer torque

rheometer (Model MTR, Caleva Ltd, Dorset, England). Dry powder was mixed in the

rheometer for 3 minutes to obtain the baseline response, which was afterwards deducted

from the torque responses. The quantity of dry powder was selected so that the mixing

blades were covered (IV, Table 1.). Water was added in a single addition and the mixture

was wet massed at 52 rpm for 12 minutes. Curve fittings (three parameter log normal)

were made by SigmaPlot 4.0 (SPSS Inc., Chicago, USA).

When measuring wet granules the rheometer was initially run empty at 52 rpm

for 15 sec in order to generate a baseline value (II). Thirty-five grams of wet powder

mass was then added to the bowl, and the rheometer was run for 30 sec, after which data

was recorded for 30 sec. In order to study the effect of mixing time, wet powder masses

were mixed further for 2.5 minutes. All the wet massing experiments were performed in

triplicate and a mean torque value and a standard deviation were calculated.

4.5 Powder rheometry (II)

An automated powder rheometer (Wet and Dry Powder Rheometer FingerPrint,

Freeman Technology, Malvern, UK) was used to study the wet massing behaviour of

Prosolv 50. A downward compaction and an upward slicing were used with the helical

rotor angle of 3°. The blade tip velocity was 100 mm s-1. Water was added to 35 g of dry

powder during the first 6 minutes of mixing using a Gilson pump (Miniplus 3, Model

M312, Villiers, France). Mixing of the wet powder mass was continued up to 15 minutes

(30 cycles).

25

When measuring wet granules a blade tip velocity of 40 mm s-1 was used and wet

powder masses were mixed for 12.5 minutes (10 cycles). The wet mass column height

was kept at 95-100 mm, which corresponds to 50-60 g of wet granules depending on the

moisture content. Experiments were performed in triplicate, and an equipment

dependent software was used to derive the coefficients of the functions for each

individual traverse.

4.6 Capillary rheometry (III)

The wet masses were extruded using a ram extruder designed by Ovenston and Benbow

(1968). When different ram velocities and dies with different lengths are used, ram

extruder can be used as a capillary rheometer. The barrel (diameter 2.54 cm) was fitted

with a hardened steel die of 1.5 mm diameter with different lengths (3-12 mm). The ram

extruder was driven downwards by a mechanical press (Lloyds MX-50, Southampton,

UK) using five different ram speeds (400, 200, 100, 50 and 25 mm min-1) with one filling

of the barrel. Extrusions were made in triplicate. The non-linear curve fittings of

mathematical models (Casson, Herschel-Bulkley and Power law) were made using the

program SigmaPlot 5.0 (SPSS Inc., Chicago, USA).

4.7 Near infrared spectroscopy (IV)

NIR spectra were measured using an FT-NIR spectrometer (Bühler NIRVIS, Uzwil,

Switzerland) with a fiber-optic probe. Diffuse reflectance spectra were measured over the

range of 1000-2500 nm (500 data points), each spectrum being the average of four scans.

The probe was gently placed into the wet powder mass. Each moisture level was

measured five times and the average of these spectra was used. The spectral treatment

(absorbances and second derivatives) was performed with NIRCAL v. 2.0. The second

derivative was calculated using the 9-point Savitzky-Golay algorithm. Two points’

baseline correction was performed with Matlab 5.3 (The MathWorks Inc., Natick, MA,

USA).

4.8 Thermoporosimetry and solute exclusion (V)

DSC measurements were carried out on a Mettler 821e DSC (Mettler Toledo,

Switzerland) equipped with an intracooler. The samples of approximately 3.5 mg were

weighed into non-hermetically sealed aluminium pans. The weight of the samples was

checked before and after the DSC measurements to ensure that the pan was well sealed.

26

The water content of the samples was determined by drying the samples for 24 h at

105°C.

4.8.1 Isothermal step melting procedure

In the beginning of the measurement the sample was cooled at 10°C min-1 from +25°C

to -45°C and then heated to -35°C, where it was hold for 2 minutes. The isothermal step

melting procedure was run at 1°C min-1 to set points shown in Table 1 (V). The duration

of isothermal segments varied from 2 min at -33°C to 22 min at -0.2°C depending on the

time the sample needed to complete the melting. In order to determine the total freezing

water (TFW), the sample was dynamically cooled at the end of the measurement to -45°C

at 10°C min-1 and then heated to +25°C at 5°C min-1. The results are the mean of four

measurements and the statistical differences were analysed with Student’s t-test.

4.8.2 Total bound water (TBW) measurements

The principle of the total bound water measurement is to raise dynamically the

temperature of a frozen sample slightly below 0°C, where it is held constant until the

melting transition is completed. With this method the pore size distribution can not be

determined, but the total amount of water with depressed melting temperature can be

measured.

TBW measurements were started from +3°C by cooling the sample first to -30°C

and then heating it to -0.2°C at 5°C min-1. The temperature was kept at -0.2°C for 12

minutes to ensure that the sample had enough time to complete the melting. The sample

was then cooled to -30°C and the amount of tbwH was calculated by integrating this

freezing peak (V,Eq. 4). In the end of the measurement the sample was heated to +25°C

to determine the tH (V, Eq. 2). The measurements for SMCC wet granules were

performed in duplicate and the average is reported. The results for ungranulated and

granulated cellulose samples are the mean of four measurements.

4.8.3 Solute exclusion technique

For the fiber saturation point (FSP) measurements, 2 g of the wet sample was dispersed

with water to obtain a 20% dispersion. Thirty-five millilitres of a 2% solution of 2 x 106

Dalton dextran polymer (T2000 from Amersham Pharmacia Biotech AB, Uppsala,

Sweden) was added to the dispersion and stirred for 1 hour. After stirring most of the

solution outside the particles was removed by centrifugation, after which the solution

27

was filtered. The concentration of dextran solution was determined using a polarimeter

(Autopol IV, Rudolph, USA).

28

5. RESULTS AND DISCUSSION

5.1 Torque measurements (I, II)

5.1.1 Mixing kinetics

When a binder liquid is added to a dry material, it is gradually distributed throughout the

powder bed by the mixing process. In mixer torque rheometer (MTR) measurements

water was added as a single bolus in the beginning of the experiment, whereas in powder

rheometer (PR) measurements water was added gradually during the first 6 minutes of

mixing. In addition, in PR measurements water was added onto the top of the powder

bed, and it was first mixed to the upper part of the powder bed. When mixing was

continued, water was distributed further through the powder bed. Provided that the

powder mass is homogenous, the torque is expected to increase as the blade height

becomes lower, i.e. as the rotor approaches the bottom of the vessel. After 10 minutes of

mixing the torque increased evenly at each water content, i.e. water was equally

distributed throughout the wet powder masses (II, Figs. 1a-c).

Due to the differences in liquid addition, water distribution through the powder

bed was slightly slower in PR as compared to MTR. However, when comparing the

torque curves of Prosolv, it can be seen that the mixing kinetics was similar in both the

rheometers. At low liquid amounts (0.8-1.2 g g-1), all the torque curves were descending

after few minutes of mixing (Fig. 9a and I, Fig. 5a). At higher amounts of liquid (1.3-1.5 g

g-1), the curves turned into ascending ones (Fig. 9b and I, Fig. 5b).

Figure 9. Mixing kinetics of Prosolv in a powder rheometer.

The results are consistent with Carstensen et al. (1976), who studied the mixing

kinetics of water using a planetary mixer. They found out that in the initial stages, part of

Mixing time (min)

0 2 4 6 8 10 12 14 16

Torq

ue (m

Nm

)

0

100

200

300

400

500

1.3 g g-1

1.4 g g-1

1.5 g g-1

b)

Mixing time (min)

0 2 4 6 8 10 12 14 16

Torq

ue (m

Nm

)

0

100

200

300

400

500

0.8 g g-1

1.1 g g-1

1.2 g g-1

a)

29

the mixture is excessively wetted and forms large aggregates. The powder which is not

associated with these aggregates, remains dry. Similarly, Alleva (1984) observed an initial

force peak at the low to intermediate liquid contents, and attributed this to the uneven

distribution of the granulating liquid and the formation of large unstable agglomerates.

On the surface of the particles the system exhibits an increased resistance because of

agglomerates, and an increased autoadhesion due to the presence of excess liquid.

5.1.2 Effect of mixing time

All the material exhibited an increase in torque with increasing water content rising to a

maximum, whereafter the torque decreased as the material became overwetted (I, II).

Alleva (1984) and Rowe and Sadeghnejad (1987) suggested that this behaviour is

consistent with the different states of liquid saturation (pendular, funicular and capillary),

as defined by Newitt and Conway-Jones (1958). The wet masses containing low amounts

of water exhibit low torque values, since the properties of the wet mass remain similar to

those of dry material until sufficient liquid bridges are formed. The overwetted masses

exhibit low torque values due to the additional water, which cannot be taken inside or

between the particles, and water acts as lubricant.

After water was homogeneously distributed to the wet mass, the maximum

torque value was achieved at 1.2 g g-1 with both the rheometers (Fig. 10). The greatest

difference between the rheometers was that the driest masses gave much higher torque

values in PR as compared to MTR. This is because PR is more sensitive to dry powders

compared to MTR. In MTR the dry powders give the same torque as the empty bowl,

whereas PR can be used to

study also the dry powders.

In both the rheome-

ters the peak torque

location shifted to higher

amounts of liquid as a

consequence of prolonged

mixing (I, Fig. 2 and II, Fig.

2). In the conventional

granulation the liquid satu-

ration expresses the volume

Liquid addition (g g-1)

0.6 0.8 1.0 1.2 1.4 1.6

Mea

n to

rque

(Nm

)

0.0

0.5

1.0

1.5

2.0

Torq

ue (m

Nm

)

0

100

200

300

400

500Torque in MTRTorque in PR

Figure 10. Comparison between mixertorque and powder rheometer.

30

of liquid relative to the volume of voids and pores between the particles in an

agglomerate. The transition between different states of liquid saturation is usually

induced by increasing the level of binder liquid, but an identical effect can be obtained

also by consolidating the agglomerates by further mixing (Newitt and Conway-Jones,

1958; Kristensen and Schæfer, 1987). However, the liquid saturation model has been

developed for non-porous and spherical sand particles, which do not have the same

ability as MCC/SMCC to swell and absorb large amounts of water into the internal

structure. In the case of MCC/SMCC, prolonged mixing will cause an increased

absorption of water, and consequently the liquid saturation of about 100% occurs at a

higher amount of liquid (I).

5.1.3 Effect of shear forces

The torque values of wet granules behaved differently in PR than the torque values after

liquid addition to the dry powder (II). The torque values of wet granules were also lower

than the values after liquid addition. It is possible that the different velocities used in

liquid addition and wet granules measurements caused the observed differences between

the torque values. However, it has been shown previously that MCC is able to

immobilise more water during granulation if high shear forces are used as compared to

low shear forces (Schmidt and Kleinebudde, 1999). Consequently, another explanation to

different torque values could be that more water is able to go into the internal structure

of microcrystalline cellulose during granulation in a planetary mixer than during water

addition experiments in the powder rheometer. Therefore, most of the liquid in the

system is present within the granules, not on the surface. This causes the lower

interparticular contacts and lower torque values.

In MTR experiments the values of the peak torque of wet masses were attained at

the same water contents both in the liquid addition and wet granule experiments (II).

This was not the case in powder rheometer experiments. However, it is notable that

different planetary mixers were used in these experiments. It is possible that the shear

forces applied to the wet powder mass were higher in the planetary mixer (Hobart) used

in PR studies as compared to the planetary mixer (Kenwood) used in MTR studies. It

seems that after granulation water is more strongly bound into the internal structure of

microcrystalline cellulose either because of higher shear forces or longer equilibrium time

as compared to the liquid addition studies.

31

5.1.4 Evaluation of methods

The mixing kinetics and the effect of mixing time proved to be similar with both the

rheometers. The fact that also dry powder masses can be studied in the powder

rheometer, which makes it more useful as compared to the mixer torque rheometer.

However, before the powder rheometer can be properly utilised in wet powder mass

studies, the problem of torque overload requires resolution.

5.2 Rheological properties obtained by capillary rheometry(III)

5.2.1 Flow curves

The steady state extrusion pressure increased with the increasing length to radius (L/R)

ratio of the die and also with the increasing ram speed (III). In general, the shear stresses

decreased with the increasing water content with all the cellulose grades (III, Fig. 2). The

results are consistent with the previous findings by Harrison et al. (1985) and Raines et al.

(1989). The derived flow curves showed a shear thinning behaviour , which indicates that

the materials realign or reorientate themselves or there is an appreciable water movement

so that there is less resistance to flow at higher shear rates (III). The water contents used

affected the rheological properties of cellulose grades slightly differently, which indicates

that the cellulose grades reacted with water in a different way.

5.2.2 Rheological models

The majority of mathematical models used in rheology express a non-linear relationship

between shear stress and shear rate. Parameters derived from the rheological models

have been used to determine the variations between different materials. From the

different rheological models evaluated, it was not possible to indicate, which model

described the rheological properties of each cellulose grade at each water level the best

(III, Table 1). The residuals were shear rate dependent, which revealed that the models

did not perfectly agree with the physical properties of the wet masses. An interesting

finding was that the yield stress of the Casson model showed an identical behaviour with

the Power Law viscosity constant (III, Figs. 3a-b). Since no measurements were made at

low shear rates, the existence of real yield stresses remained unclear.

32

5.2.3 Elastic properties

The elastic properties of wet masses increased with the increasing water content and

decreased with the increasing shear stresses (III, Figs. 5a-c). The SMCC grade proved to

be more elastic than the MCC grades at each moisture content. Thus, the elastic

properties of MCC and SMCC wet masses were different and changed with water

content. Consequently, it is not possible to achieve similar rheological properties between

the different grades of cellulose by altering the water content of the wet mass.

It has been suggested that there appears to be an optimal elasticity for the

formation of satisfactory spheres in the spheronisation stage (Chohan and Newton,

1996). Wet powder masses with high elasticity or low plasticity may resist the

spheronisation by restoring the original shape. Chohan and Newton (1996) and El Saleh

et al. (2000) found out that relatively inelastic materials, such as the colloidal grades of

MCC, tend not to round during the spheronisation. However, the silicification of MCC

did not affect its behaviour during pellet production (El Saleh et al., 2000). Since the

elasticity of colloidal grades were of the same magnitude as those reported for Avicel,

Emcocel and Prosolv in this study, there has to be also some other properties than

elasticity, which affect moulding of material during spheronisation.

5.2.4 Evaluation of capillary rheometry

Even though it is possible to use capillary rheometer at different speeds with a single

barrel of wet mass, the measurements are time-consuming. In addition, the equilibrium

time of the wet mass and the material of the die (hardened vs. stainless steel) affect the

results (Raines, 1990). It is also notable that capillary rheometer does not give any

information about the properties of wet granules but the rheological properties of the

material. This is because the wet granules are densified during the compression stage of

extrusion, and the possible differences in liquid saturation are diminished. In order to

compare different materials and describe the rheological properties precisely, one should

find a rheological model, which describes the material well. This was not achieved in this

study. The reason why the rheological models did not fit to the flow curves of

MCC/SMCC masses is probably the complex structure of cellulose compared to e.g.

melted polymers.

33

5.3 Near infrared spectroscopy (IV)

5.3.1 Absorbance spectra

The NIR spectra of dry

cellulose and lactose showed

similar features, since carbo-

hydrates mainly include CH

and OH bonds (Fig. 11). The

major absorption band of

water in the NIR region is at

around 1940 nm (Curcio and

Petty, 1951; Osborne et al.,

1993). The water of SMCC

was seen as a wide band,

whereas the hydrate water of

lactose was seen as a narrow

water band. The shape of the

water band depends on the

changes in the frequencies

due to hydrogen bonding

occurring within the environ-

ment where the water is

located (Osborne et al.,

1993). The broad band for

adsorbed water of SMCC

indicates a spread of energies

of interaction, whereas the

mono-hydrate water band is

typical to a more uniform

interaction. The absorbance

spectrum of the 1:1 mixture

was a combination of the

SMCC and lactose spectra.

1000 1200 1400 1600 1800 2000 2200 2400-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

Wavelength (nm)

Abso

rban

ce, l

og10

(1/R

)

SMCC Lactose 1:1 mixture Glass ballotini

a)

1350 1375 1400 1425 1450 1475 1500

-1.5

-1

-0.5

0

0.5

1

x 10-3

Wavelength (nm)

2nd

deriv

ative

of l

og10

(1/R

)

SMCC Lactose 1:1 mixture Glass ballotini

Monohydrate

b)

1825 1850 1875 1900 1925 1950 1975 2000

-8

-6

-4

-2

0

2

4x 10-3

Wavelength (nm)

2nd

deriv

ative

of l

og10

(1/R

)

SMCC Lactose 1:1 mixture Glass ballotini

Monohydrate

c)

Figure 11. a) NIR absorbance spectra, b)Second derivative spectra of the overtonewater band and c) Second derivative spectra ofthe combination water band of dry materials.

34

Glass ballotini showed no absorbance, since it is an inorganic material.

The water bands at 1450 and 1930 nm were increased with the increasing

moisture content for all materials (IV, Figs. 2a-d). Since the physical changes in the

samples affect the NIR spectrum, an upward displacement of the absorbance spectra

baseline was observed with an increasing moisture content as well. When particle size

increases, the radiation will penetrate deeper into the sample and a decrease in the back

reflected light is seen as an increase in the apparent absorbance (log(1/R)). Similarly, the

presence of water affects the overall spectrum, since the scattering depends on the ratio

of the refractive index of the particles, as compared that of the surrounding medium

(Osborne et al., 1993). This means that the NIR spectroscopy is a surface method, i.e.

NIR probe analyses the amount and the state of water on the surface of the particles.

Due to the porous structure of cellulose, SMCC is capable of absorbing large

amounts of water into the internal structure. The water in the cellulose samples has been

attributed to the bulk water between fibers and to the water in pores within the fiber (Li

et al., 1992). In addition to the water in the pores and at the surfaces, water is also

absorbed into the amorphous parts of cellulose. In the case of non-porous material, such

as glass ballotini, water can only be adsorbed onto the surface of particles. Due to the

different refractive properties of the wet masses, the upward displacement of the

absorbance spectra baseline was the greatest with glass ballotini and α-lactose

monohydrate, even though the moisture contents were much lower as compared to

SMCC (IV, Figs. 2b, 2d). The baseline corrected heights of the water bands (moisture

curves) at 1933 nm for all the four materials are shown in Fig. 12. The moisture curves of

glass ballotini and α-lactose

monohydrate rose up very fast

and almost linearly. As SMCC is

capable of holding large

amounts of water in the internal

structure, the same amount of

moisture resulted in higher water

bands with other materials at

moisture levels from 0.1 ml g-1

and up.0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Moisture content (ml g-1)

Bas

elin

e co

rrect

ed a

bsor

banc

e

SMCC 1:1 mixture Lactose Glass ballotini

Figure 12. Effect of moisture content on theheight of the baseline corrected water bands at1933 nm.

35

5.3.2 Second derivative spectra

A common method of correcting the baseline of the spectra is to use the second

derivative. The minimum in the second derivative spectra is related to the maximum in

the absorbance spectra. The tightly bound water of SMCC was seen at 1425 and 1920

nm, whereas the hydrate water of α-lactose was seen at higher wavelengths (1450 and

1933 nm) (Fig. 11). The absorption maximum of water band is dependent on the

hydrogen bonding environment of the water (Osborne et al., 1993). Since the hydrate

water includes more hydrogen bonds than the water of cellulose, it is shown at higher

wavelengths. The second derivative spectra of the 1:1 mixture were capable of separating

the tightly bound and monohydrate water bands, which overlapped in the absorbance

spectrum (Fig. 11). The sharp monohydrate bands of the 1:1 mixture were seen at the

same wavelengths (1450 and 1933 nm) as with the lactose.

With increasing mois-

ture content the water bands of

SMCC increased in size and

shifted gradually from 1920 nm

to 1903 nm (IV, Fig. 6a). The

shift of the second derivative

band maximum is shown also

in Figure 13. This shift to

lower wavelengths indicates an

increase in the median inter-

action energy of the water -OH

bond, since non-hydrogen-

bonded -OH groups have higher stretch frequencies and, hence, a lower wavelength than

the hydrogen-bonded -OH groups. The second derivative spectra of α-lactose

monohydrate were capable of distinguishing the hydrate water from added water (IV,

Fig. 6b). Added water resulted in a separate band at a lower wavelength (1903 nm) in

comparison with the structured monohydrate band at 1933 nm. The water band of the

1:1 mixture was seen at a slightly lower wavelength (1910 nm) than that of pure SMCC

(1920 nm) and, consequently, a smaller shift from 1910 nm to 1903 nm was observed

(IV, Fig. 6c). This is due to the closeness of the monohydrate and cellulose water bands.

The sharp monohydrate band at 1933 nm remained unchanged after water addition. In

Figure 13. The location of maximumsecond derivative band height.

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6-7

-6

-5

-4

-3

-2

-1

0

1 x 10-3

Moisture content (g g-1)

2nd

deriv

ative

ban

d he

ight

1920 nm1910 nm1900 nm

36

the case of glass ballotini, the water bands were seen directly at 1903 nm (IV, Fig. 6d).

The intensity of the water bands increased with the increasing water content, while no

shift in the wavelength could be observed.

5.3.3 Evaluation of NIR spectroscopy

The second derivative spectra should be very useful in detecting pseudopolymorphic

changes, since it is capable of distinguishing the hydrate water from the other energetic

states of water (IV). In fact, NIR spectroscopy has been used to study the

pseudopolymorphic transformations of lactose and theophylline (Buckton et al., 1999;

Räsänen et al., 2001). Rantanen et al. (2000b) observed that the water band of

theophylline increased non-linearly due to the pseudopolymorphic changes during the

granulation. Comparison of the ratio of the second derivative spectra at two wavelengths

enables the detection of low levels of one polymorphic form in the presence of another

(Luner et al., 2000; Seyer et al., 2000).

The upward displacement of baseline and the relative height of water bands were

the greatest with materials that had a poor liquid retention capacity (IV). Consequently,

the energetic state of water in the wet mass depends on the physical and chemical

properties of the material. This enables the study of the liquid retention capacity of

different materials using NIR. The results agree with Watano et al. (1996) and Miwa et al.

(2000). The authors used NIR and water absorbing potential of materials to predict the

suitable amount of water for wet granulation. However, these results were in contrast to

the results obtained from fluidised bed granulation process (Rantanen et al., 2001). Due

to the low shear forces during the fluid bed granulation, less water is able to penetrate

into the internal structure of microcrystalline cellulose as compared to granulation in

planetary mixers. Thus, when evaluating solid-water interactions with NIR, the effect of

shear forces on the state of water should be considered.

5.4 Thermoporosimetry (V)

5.4.1 Different water fractions in SMCC wet masses

At low moisture contents (0.05-0.2 g g-1) all the water was bound on the surface of

cellulose, i.e. water was nonfreezing (Fig. 14). The amount of nonfreezing water

increased with the increasing moisture content up to 0.26 g g-1. The results are consistent

with the amounts of nonfreezing water found for starches and pulp samples (Mousseri et

37

al., 1974; Nelson, 1977; Hatakeyama et al., 1983). There is still not a well-accepted

thermodynamic theory to explain why some of the water in the hydrated systems does

not freeze. However, it has been shown that there is a correlation between the number

and type of available hydration sites and the quantity of nonfreezing water (Nakamura et

al., 1981; Berthold, 1996). Nonfreezing water is often divided to the tightly bound

monolayer water and to the intermediately bound multilayer water, which corresponds to

about three water layers next to the surface (Zografi and Kontny, 1986; Li, 1991;

Delwiche et al., 1991).

The sum of nonfreezing water (NFW) and freezing bound water (FBW) is

referred to as the total bound water (TBW), which shows the amount of water in the

pore range of 1-216 nm (Fig. 14). The amount of TBW increased with the increasing

moisture content and reached a plateau at 1.1 g g-1. Free water, which cannot be detected

by DSC alone, is located in larger pores (> 216 nm). It is probable that the pores of the

wettest granules are larger than

216 nm and they cannot be

separated from the free water with

thermoporosimetry. Fiber satura-

tion point (FSP) measures the

total amount of water inside the

cellulose particles, i.e. swelling of

the particles. When the moisture

content of the sample exceeds the

FSP, the additional water is

retained between the particles as

bulk water.

5.4.2 Effect of granulation

Granulation had no effect on the amount of nonfreezing water, since it was the

same for the ungranulated and granulated wet masses (V, Table 2). With the

ungranulated wet masses the amounts of NFW and FBW were almost equal, and only a

very small amount of water was in the larger pores as free water. The FSP values of

ungranulated wet masses were between 0.41 and 0.46 g g-1 and, hence, most of the added

water was as bulk water outside of the particles. With the granulated wet masses the

amount of FBW was twice as much as the amount of NFW and free water. In general,

Moisture content (g g-1)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

Pore

wat

er (g

g-1

)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

NFW

TBW

Nonfreezing water

Freezing bound water

Free water

Bulk water

FSP

Figure 14. Different water fractions ofSMCC wet granules.

38

the granulated wet masses contained no bulk water, since the FSP values of wet granules

were very close to the moisture content of wet granules. Consequently, all of the water

was inside the particles. Most of the water was in the small pores (TBW) and smaller part

in the larger pores (free water). The differences between water fractions of ungranulated

and granulated wet masses were also statistically significant (V, Table 3).

Granulation seemed to have the same effect on the microcrystalline cellulose as

beating has on cellulose fibers in the papermaking process. Beating is known to cut the

cellulose fibers and cause the morphological changes. The primary effect of beating on

the cellulose is to open larger pores in the cell wall, and hence increase the swelling of the

pulp fibers (Stone et al., 1968). Consequently, the lamellar structure of fibrils inside the

cellulose fibers is loosened which increases the water penetration into the cell (Emerton,

1980; Li, 1991). It is expected that the granulation process works in the same way; bonds

between microfibrils are loosened and the particles swell.

As Froix and Nelson (1975) stated, the amount of water contained in a cellulose

system is a direct function of the microstructure of that system. Thus one should not

expect to obtain identical results with the different samples without normalisation of

water accessibility into the structure. Mixing time and shear forces applied to the wet

mass during granulation affects the location of water in the MCC/SMCC wet mass, i.e.

the water fraction inside/outside the particles. This means that one should not compare

samples without taking account whether the wet mass is granulated or not, and which

granulation method has been used.

5.4.3 Evaluation of thermoporosimetry

The pore structure of dry microcrystalline cellulose powder has been traditionally

determined either by using nitrogen adsorption or mercury porosimetry where the

sample is evacuated before the measurements (Nakai et al., 1977; Hollenbeck et al., 1978;

Zografi et al., 1984). The pore size range obtained by thermoporosimetry (1-216 nm) is

quite close to that obtained by nitrogen adsorption (0.3-300 nm). When the pore size

distribution of dry cellulose samples were measured by nitrogen adsorption, almost all of

the pores were under 100 nm (I). The result is consistent with Westermarck et al. (1999).

However, when the pore size distribution of wet cellulose samples was measured by

thermoporosimetry, the pore size of the samples increased with the increasing water

content. In general, there were more water in the pore range of 100-216 nm as compared

to smaller pores.

39

In the interaction with water the cellulose particles swell generating new pores,

but during drying most of the pores collapse. Consequently, in order to measure the

pores of wet MCC, the pore structure should be measured with the techniques directly

applicable to the water saturated particles. In this study the pore size distribution of wet

MCC/SMCC powder masses was measured using thermoporosimetry. In addition,

thermoporosimetry was able to differentiate the freezing bound water from the free/bulk

water, which normally overlap in the melting endotherms. Both the step melting

procedure and the TBW method proved to be suitable for that purpose.

Freezing the sample is an essential part of the thermoporosimetry. However,

crystallisation of water within the cell wall may affect the structure of the sample and,

hence, the measured pore size distribution. In fact, Maloney et al. (1998) noted

considerable hysteresis if the freezing/melting cycle was repeated several times. With the

above methods hysteresis can be minimised since the sample is frozen and melted only

twice. While thermoporosimetry has been shown to be in good agreement with the other

pore measurement techniques for porous glasses, this has not been shown for the pulp

fibers (Maloney, 1999). Thus it is likely that the relationship between the pore size and

the melting temperature in the pulp fibers is more complicated than described by the

Gibbs-Thomson equation (Maloney et al., 1998). An other disadvantage of the method is

that the pore range, which can be measured is fairly narrow (1-216 nm). In addition, the

sample sizes used are so small that it is quite difficult to get a representative sample from

the wet powder mass. However, even though it may be difficult to ascertain the actual

size of the pores of MCC/SMCC wet masses with thermoporosimetry, the method can

be used to measure the different water fractions in wet powder masses and to make

relative comparisons between the different samples.

5.5 The difference between cellulose grades

5.5.1 SMCC vs. MCC grades

The mean diameter of all cellulose particles varied from 58.6 to 59.7 µm, and there were

no significant differences between the particle size distributions of cellulose grades either

(I). The SEM figures showed that also the shape of the particles was similar between the

different grades (not shown).

The specific surface area of the SMCC grade was around five times greater than

those of the MCC grades (I). The result was associated with the large specific surface area

40

of colloidal silicon dioxide. In addition, the SMCC grade showed better flowability and

greater bulk and tapped densities than the MCC grades. The differences in the densities

are probably caused by the improved flowing and packing characteristics of the SMCC

grade.

The relative increase in swelling was greater for Prosolv as compared to Avicel

and Emcocel (I). The same phenomenon could be seen also in the contact angle

measurements. Prosolv tablets swelled very fast and the measurement had to be ended

much before those of the MCC grades. In addition, the contact angle of Prosolv was

significantly lower than those of Avicel and Emcocel (Fig. 15). During swelling the

cellulose particles absorb water into the internal structure, which means that there is less

water between the cellulose particles. This caused lower liquid saturation and,

consequently, lower torque values in the MTR experiments (I) and higher shear stresses

in the capillary rheometer (III) at the lowest moisture content studied (0.8 ml g-1).

Water is essential in the

paper making process, since the

cellulose fibers do not form

hydrogen bonds to the

neighbouring cellulose fibers

without water. During drying

cellulose-water hydrogen bonds

between fibers are replaced by

cellulose-cellulose hydrogen

bonds, which give paper its

strength. Some of these hydrogen

bonds might not open again when

cellulose is reintroduced to water. The phenomenon is called hornification in the pulp

and paper literature according to which term quasi-hornification has been used in

pharmacy (Staniforth and Chatrath, 1996). Even though the greater swelling of Prosolv

had quite small influence at higher water contents, it might have practical meaning in the

traditional granulations, which are performed at around the lowest moisture content used

in this study (0.8 ml g-1)(I-II). If Prosolv is able to absorb more water into the internal

structure, there is less water between the cellulose particles, which might reduce the

hydrogen bond formation between the cellulose particles. In fact, SMCC has been

claimed to be more resistant to the wet granulation than the conventional MCC grades,

Figure 15. Contact angles of differentcellulose grades as a function of time.

Time (s)

0.0 0.2 0.4 0.6 0.8 1.0

Con

tact

ang

le (°

)

0

5

10

15

20

25

30

35

Avicel PH 101Emcocel 50Prosolv 50

41

since SMCC maintains compaction properties after wetting and drying (Sherwood and

Becker, 1998). In addition, Sherwood and Becker (1998) and Habib et al. (1999) found

out that the decrease in the granule porosity was more pronounced for the MCC grades

as compared to the SMCC grade. Habib et al. (1999) observed that the decrease in

granule porosity was more pronounced for both the materials after 0.8 g g-1. The authors

attributed the effect of high water levels to the greater intragranular interaction between

the neighbouring MCC/SMCC particles. Consequently, the results are consistent with

the swelling results obtained in this study.

Li et al. (2000) suggested that the physical reason for the difference between the

MCC and SMCC wet masses may be caused by the presence of silicate particles that

strengthen the interparticulate interactions. However, more probable reason for the

differences seen in their study is the greater swelling of SMCC grade, since the moisture

level used was close to our lowest moisture content (1.0 g g-1). Strong interparticulate

interactions has been suggested to cause also the stronger tensile strength of SMCC

tablets in the direct compression (Edge et al., 2000). The mechanism behind this

behaviour is still unknown, but it is presumable caused by the large specific surface area

of SMCC, which increases the contact points between the particles.

The different surface properties of dry SMCC as compared to MCC grades are

well documented, i.e. the better flowability and the large specific surface area. According

to the results obtained in this study the surface properties of the SMCC grade are

different also in the wet stage. With the powder rheometer the water addition

experiments could be done only with Prosolv, because an overload occurred with Avicel

and Emcocel at those water contents, which caused the highest measurable torque values

to be exceeded (II). In the capillary rheometer studies the SMCC grade proved to be

more elastic than the MCC grades at each moisture content (III). Consequently, it seems

that the wet SMCC flows better as compared to the MCC grades.

Although the silicification process affected the swelling properties, and

furthermore, the mixing kinetics of microcrystalline cellulose, the source of the

microcrystalline cellulose had a stronger influence than silicification on the liquid

requirement at the peak torque (I). The silicification does not induce gross changes in the

shape or texture of the microcrystalline cellulose particles (Tobyn et al., 1998) and,

consequently, the large specific surface area had no influence on the liquid requirement

during the torque experiments. The results are consistent with Zografi (1988) who stated

42

that in case the material is capable of absorbing water into the solid structure, water

uptake is independent of the specific surface area.

5.5.2 MCC grades

The properties of the MCC grades were almost identical when the dry powders were

studied. The particle size and size distribution, density, flowability and specific surface

area of Avicel and Emcocel were similar (I). In addition, the relative increase in swelling

and the contact angles were alike.

Emcocel wet masses exhibited higher torque values at a lower water content than

Avicel masses (I, II). This indicated that Emcocel masses achieved the liquid saturation

of about 100% at a lower liquid amount than Avicel masses. This means that Emcocel

masses were not able to hold as much water in the internal structure as Avicel. When the

different water fractions were measured using thermoporosimetry, it was noticed that the

amount of freezing bound water (FBW) was slightly lower with Emcocel as compared to

Avicel and Prosolv. Thermoporosimetry is able to measure only the relatively small pores

(1-216 nm), which means that the FBW is located in this pore range. The smaller amount

of FBW indicates that Emcocel has less small pores, and thus a greater amount of larger

pores as compared to Avicel and Prosolv. Thus, the different pore structure of wet

cellulose masses might be one explanation for the differences seen in the torque

measurements.

Consequently, the difference between the MCC grades was seen only when wet

massing was included to the characterisation, which indicates that their interaction with

water is different. The results agree with those obtained by Raines (1990) and Newton et

al. (1992). Raines (1990) observed that the various grades of MCC have different

rheological properties when mixed with water and lactose. Newton et al. (1992) showed

that in the formulation of spherical particles by extrusion-spheronisation it is not possible

to interchange the brands of MCC without changing the water content. Thus grades,

which are considered equivalent, do not function always in an equivalent manner.

Several authors have studied the differences between the regular microcrystalline

cellulose grades, Avicel PH 101 and Emcocel (Staniforth et al., 1988; Raines, 1990;

Newton et al., 1992; Landín et al., 1993a,b; Kleinebudde, 1997). Major causes of the

intermanufacturer variability among microcrystalline celluloses have been associated to

the variability of the source of pulp and of the manufacturing process (Landín et al.,

43

1993a,b). Wu et al. (2001) studied the effects of the manufacturing factors on the

properties and functionality’s of the MCC products. They observed that the most

important factor in the manufacturing was the temperature, whereas the molecular mass

and particle morphology had the greatest effect on the material properties.

It is notable that only one batch of each cellulose grade was investigated in this

study. Consequently, one should not make general conclusions about the properties of

different cellulose grades, but rather notice that there exist some differences between the

cellulose grades and probably between the different batches. The material can be

different also in the different stages of the manufacturing process.

5.6 Comparison of different methods

5.6.1 Capillary rheometry vs. MTR

Some authors have reported that the data generated by the MTR compare favourably

with those measured by the capillary rheometry (Landín et al., 1995; Rowe, 1995; Rowe,

1996). Rowe (1995) observed that with the increasing water content both the yield stress

and kinematic viscosity constant increased. However, in the capillary rheometer studies

the viscosity constant (k) decreases when the water content is increased. Consequently,

their results were in contrast to the results obtained in this (III) and the other studies by

the capillary rheometer (Raines et al., 1989; Harrison et al., 1985; Fielden et al., 1992). In

the MTR experiments material exhibits an increase in the torque with the increasing

cohesiveness/tensile strength of the wet mass. When material becomes overwetted, only

the surface tension holds the wet mass together. Consequently, the cohesiveness and the

torque values decrease. This indicates that the MTR is able to measure only the liquid

saturation of the wet mass, not the classical rheological parameters, such as yield stress or

viscosity constants (III). The source of this difference is presumably to be found in the

different measuring systems included. In the case of the capillary rheometer, the resistant

to movement involves two components; the convergent flow from the wide barrel to the

narrow capillary and the flow through the capillary. In the capillary rheometer the air is

expelled during the compression stage, and the possible differences in the liquid

saturation are diminished. Thus, the material is in a very different state in these two

systems and one would not expect an identical behaviour.

MTR has been traditionally used in the wet granulation studies, whereas the

information obtained by the capillary rheometer has been related to the extrusion.

44

However, correlation of the rheological parameters to the extrusion properties is

difficult. It is particularly difficult if some other than ram extruder is used in the

production of pellets. If the shear forces and the principal of action between extruders

are different, for example radial screen vs. ram extruder, the capillary rheometer does not

necessarily give information about the extrusion properties of the material. Since the

possible differences in the pore structure of wet cellulose masses are diminished during

the compression stage of extrusion, the different grades of cellulose cannot be

differentiated as clearly as e.g. with the torque measurements. In fact, in the case of the

radial screen extruder, the peak torque in MTR measurements can be used to predict the

optimum water content for extrusion-spheronisation (Luukkonen et al., 1999).

5.6.2 NIR vs. MTR

In the case of non-porous material, such as glass ballotini, water can only be adsorbed

onto the surface of particles. With these materials the peak torque values were attained at

very low water contents (0.03-0.14 ml g-1). When half of the α-lactose monohydrate mass

was replaced with SMCC, the torque values were considerably higher and the peak torque

values were achieved at higher water contents (0.60-0.65 ml g-1). The peak torque values

of SMCC were attained at the highest water contents (1.1-1.3 ml g-1), which shows that

the water content required to achieve the capillary state is strongly related to the amount

of SMCC in the wet powder mass. The results can be explained by the porous and partly

amorphous structure of SMCC as compared to the other materials. If water is adsorb

only onto the surface of the particles, the voids between the particles are filled very fast.

After this water acts as a lubricant and the cohesiveness between the particles decrease

and, thus, the torque values decrease.

In the NIR spectroscopy the upward displacement of the baseline and the

relative height of the water bands were greatest with the materials that had a poor liquid

retention capacity. With increasing moisture content changes in the physical properties of

the samples resulted in an apparent increase in log (1/R) throughout the whole NIR

spectrum. The plateau stage of baseline corrected water band heights could be correlated

to the maximum cohesiveness of the wet mass in MTR measurements (IV).

Even though NIR spectroscopy and MTR are dissimilar methods, the results

obtained by these methods do show some similarity (IV). It has been shown previously

that the ability of MCC to hold water in the internal structure depends on the shear

forces applied to the wet mass (Baert et al., 1992). This could partly explain the observed

45

differences between these two methods, since unlike MTR, NIR measurements do not

involve mixing of the wet mass. The results suggest that NIR spectroscopy could be used

for the end-point detection in the real time process measurements.

5.6.3 Thermoporosimetry vs. NIR

At low moisture contents (0.05-0.15 g g-1) all the water was nonfreezing (Fig. 14) and the

maximum second derivative band was at 1920 nm (Fig. 13). At moisture contents from

0.2 to 0.5 g g-1 part of the water was freezing bound water and the second derivative

maximum in NIR spectrum was at 1910 nm. Since non-hydrogen-bonded -OH groups

are obtained at lower wavelengths than hydrogen-bonded -OH groups, water was not

that tightly bound to the cellulose as compared to the lowest moisture contents. When

more water was added to the wet mass, the amount of freezing bound water increased

and the second derivative maximum shifted to even lower wavelength (1900 nm). At

high moisture contents (1.0-1.4 g g-1) the freezing bound water and the second derivative

maximum (1900 nm) achieved a steady state. It is notable that even though the maximum

band height shifted to lower wavelengths, the heights of 1920 and 1910 nm water bands

increased up to 0.8 g g-1 and to 1.0 g g-1, respectively.

These two methods can be evaluated also by comparing the baseline corrected

absorbance band heights of NIR spectroscopy to the amount of freezing bound water

(Fig. 16). Both the band height in the NIR experiments and the amount of TBW in

thermoporosimetry measurements increased with the increasing moisture content up to

1.0 g g-1 after which they achieved quite similar plateau level. These results indicate that

the changes in the state of

water in SMCC wet masses

can be seen in a similar

manner with both the

methods.

The plateau stage of

SMCC water band heights

observed in the NIR

measurements was corre-

lated to the maximum

cohesiveness of the wet

Moisture content (g g-1)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

Base

line

corre

cted

abs

orba

nce

0.0

0.2

0.4

0.6

0.8

1.0

Pore

wat

er (g

g-1

)

0.0

0.2

0.4

0.6

0.8

1.0

Total bound water (TBW)Height of water band in NIR

Figure 16. Comparison of NIRspectroscopy and thermoporosimetry.

46

mass in the MTR measurements (IV). The maximum torque values have been previously

connected to the maximum cohesiveness of the wet mass (Leuenberger et al., 1979), to

the capillary state of the liquid saturation (Alleva, 1984; Rowe and Sadeghnejad, 1987)

and to the optimum water content in extrusion-spheronisation (Miyake et al., 1973;

Souto et al., 1998; Luukkonen et al., 1999).

5.6.4 Applicability of methods

Torque measurements and NIR spectroscopy are suitable for all the materials and, above

all, they would be suitable also for the monitoring of real time processes. Since NIR

spectroscopy is able to detect both the physical and chemical information from the

sample, it is a more tempting choice for the process control. The torque measurements

are fast to perform and, consequently, they would be suitable for the product

development. The torque measurements are also able to predict the water amount

needed for extrusion-spheronisation, at least for the radial screen extruder. The capillary

rheometry can be used for studying the rheological properties of material and for

screening of materials suitable for extrusion-spheronisation. However, the method is

slow and the correlation of the rheological parameters to the extrusion properties is

difficult. Thermoporosimetry combined with the solute exclusion technique is able to

explain why the wet masses behave differently in the torque measurements. Both

thermoporosimetry and solute exclusion technique demand sample collection and are too

slow for the process control, but the methods can be used for the characterisation of new

materials. Disadvantage of thermoporosimetry and solute exclusion technique is that they

are not suitable for the water-soluble materials. Consequently, the methods used in this

study support each other and together they give extensive information about the

properties and behaviour of the wet cellulose masses.

47

6. CONCLUSIONS

SMCC grade exhibited improved flowability and increased specific surface area as

compared to the MCC grades. The differences between the MCC grades were seen only

when wet massing was included to the characterisation. Although the silicification

process affected the swelling properties, and furthermore, the mixing kinetics of

microcrystalline cellulose, the source of the microcrystalline cellulose had a stronger

influence than silicification on the liquid requirement at the peak torque.

The mixing kinetics and the effect of mixing time proved to be similar with the

mixer torque and powder rheometer. The fact that also the dry powders can be studied in

the powder rheometer makes it more useful as compared to the mixer torque rheometer.

However, before the powder rheometer can be properly utilised in the wet powder mass

studies, the problem of torque overload requires resolution.

The rheological properties of MCC and SMCC wet masses obtained by capillary

rheometer were different and changed with the water content. The SMCC grade proved

to be more elastic than the MCC grades at each moisture content. Correlation of the

rheological parameters to the extrusion properties was difficult. The possible differences

in the liquid saturation were diminished during extrusion. Thus, the different grades of

cellulose could not be differentiated as clearly as with the torque measurements.

The energetic state of water in the wet mass depended on the physical and

chemical properties of the material, which enabled the study of the liquid retention

capacity of different materials using near infrared spectroscopy. The second derivative

spectrum was capable of distinguishing the hydrate water from the other energetic states

of water. The plateau stage of baseline corrected water band heights could be correlated

to the maximum cohesiveness of the wet mass in the MTR measurements. The results

suggest that the NIR spectroscopy could be used for the end-point detection in the real

time processes.

Thermoporosimetry was able to differentiate the freezing bound water from the

free/bulk water, which normally overlap in the melting endotherms and, consequently,

four distinct fractions of water could be detected in the MCC/SMCC wet masses. Both

the step melting procedure and the total bound water (TBW) method proved to be

suitable for that purpose. The pore volume results were consistent with the NIR

spectroscopy results, which indicates that the changes in the state of water in the

cellulose wet masses can be seen in a similar manner with both the methods.

48

REFERENCES

Alleva, D.S., 1984. The wet granulation process: Kinetics and mechanisms of granule formation.Doctoral Thesis, Philadelphia College of Pharmacy & Science, USA.

Antti, H., 1999. Multivariate characterization of wood related materials. Doctoral Thesis,University of Umeå, Sweden.

Baert, L., Fanara, D., de Baets, P. and Remon, J.P., 1991. Instrumentation of a gravityfeed extruder and the influence of the composition of binary and ternary mixtures on theextrusion forces. J.Pharm.Pharmacol. 43 745-749.

Baert, L., Remon, J.P., Knight, P. and Newton, J.M., 1992. A comparison between theextrusion forces and sphere quality of a gravity feed extruder and a ram extruder.Int.J.Pharm. 86 187-192.

Bagley, E.B., 1957. End corrections in the capillary flow of polyethylene. J.Appl.Phys. 28624-627.

Barlow, C.G., 1968. Granulation of powders. Chem.Eng.(Lond.) 220 CE196-201.

Barrett, E.P., Joyner, L.G. and Halenda, P.P., 1951. The determination of pore volumeand area distributions in porous substances. I. Computations from nitrogen isotherms.J.Amer.Chem.Soc. 73 373-380.

Benbow, J.J. and Bridgwater, J., 1993. Paste Flow and Extrusion. Clarendon Press, Oxford,153 pp.

Benbow, J.J., Oxley, E. and Bridgwater, J., 1987. The extrusion mechanics of pastes – theinfluence of paste formulation on extrusion parameters. Chem.Eng.Sci. 42 2151-2162.

Berthold, J., 1996. Water adsorption and uptake in the fibre cell wall as affected by polar groups andstructure. Doctoral Thesis, Royal Institute of Technology, Stockholm, Sweden.

Berthold, J., Olsson, R. and Salmen, L., 1998. Water sorption on hydroxyl and carboxylicacid groups in carboxymethyl cellulose (CMC) studied with NIR-spectroscopy. Cellulose 5281-298.

Bier, H.P., Leuenberger, H. and Sucker, H., 1979. Determination of the uncriticalquantity of granulating liquid by power measurements on planetary mixers. Pharm.Ind. 41375-380.

Blair, T., Buckton, G., Beezer, A. and Bloomfield, S., 1990. The interaction of varioustypes of microcrystalline cellulose and starch with water. Int.J.Pharm. 63 251-257.

Boutell, S.L., 1995. Factors influencing the preparation of spherical granules byextrusion/spheronisation. Doctoral Thesis, University of London, UK.

Buckton, G., Yonemochi, E., Hammond, J. and Moffat, A., 1998. The use of near infra-red spectroscopy to detect changes in the form of amorphous and crystalline lactose.Int.J.Pharm. 168 231-241.

49

Buckton, G., Yonemochi, E., Yoon, W.L. and Moffat, A.C., 1999. Water sorption andnear IR spectroscopy to study the differences between microcrystalline cellulose andsilicified microcrystalline cellulose before and after wet granulation. Int.J.Pharm. 181 41-47.

Bugay, D. and Williams, A., 1995. Vibrational spectroscopy. In Physical characterization ofpharmaceutical solids. Brittain, H. (Ed.), Marcel Dekker Inc., New York, pp. 59-91.

Byrn, S., Pfeiffer, R. and Stowell, J., 1999. Solid-state chemistry of drugs. 2nd Ed., SSCI Inc.,USA, 576 pp.

Capes, C.E., 1980. Particle size enlargement. In: Handbook of Powder Technology, Vol 1,Williams, J.C. and Allen, T. (Eds.), Elsevier, Amsterdam, pp. 23-31.

Carstensen, J.T., Lai, T., Flickner, D., Huber, H.E. and Zoglio, M.A., 1976. Physicalaspects of wet granulations I: Distribution kinetics of water. J.Pharm.Sci. 65 992-997.

Chatlapalli, R. and Rohera, B.D., 1998. Rheological characterization of diltiazemHCl/cellulose wet masses using a mixer torque rheometer. Int.J.Pharm. 175 47-59.

Chatrath, M., 1992. The effect of wet granulation on the physico-mechanical properties ofmicrocrystalline cellulose. Doctoral Thesis, University of Bath, UK.

Chen, Z., Ikeda, K., Murakami, T. and Takeda, T., 2000. Drainage phenomenon ofpastes during extrusion. J.Mat.Sci. 35 2517-2523.

Chohan, R.K. and Newton, J.M., 1996. Analysis of extrusion of some wet powder massesused in extrusion/spheronisation. Int.J.Pharm. 131 201-207.

Choppin, G. and Downey, J., 1972. Near infrared studies of the structure of water. IV.Water in relatively nonpolar solvents. J.Chem.Phys. 56 5899-5904.

Curcio, J.A. and Petty, C.C., 1951. The near infrared absorption spectrum of liquid water,J.Optic. Soc.Am. 41 302-304.

Delalonde, M., Bataille, B., Baylac, G, Maurice, J. and Sabatier, R., 1997. Definition ofindices for the mechanical design of wet powders – application to the study of a naturalpolymer, microcrystalline cellulose. Int.J.Pharm. 146 159-165.

Delalonde, M., Bataille, B., Baylac, G. and Jacob, M., 2000. Rheological study of wetpowder masses: application to a pharmaceutical grade of microcrystalline cellulose.S.T.P.Pharma Sciences. 10 205-209.

Delwiche, S., Pitt, R. and Norris, K., 1991. Examination of starch-water and cellulose-water interactions with near infrared (NIR) diffuse reflectance spectroscopy.Starch/Stärke 43 422-427.

Dyke, T., Mack, K. and Muenter, J.S., 1977. The structure of water dimer from molecularbeam electric resonance spectroscopy. J.Chem.Phys. 66 498-510.

Edge, S., Potter, U., Steele, F., Tobyn, M., Chen, A. and Staniforth, J., 1999. The locationof silicon dioxide in silicified microcrystalline cellulose. Pharm.Pharmacol.Commun. 5 371-376.

50

Edge, S., Steele, F., Chen, A., Tobyn, M. and Staniforth, J., 2000. The mechanicalproperties of compacts of microcrystalline cellulose and silicified microcrystallinecellulose. Int.J.Pharm. 200 67-72.

Ek, R. and Newton, J.M., 1998. Microcrystalline cellulose as a sponge as an alternativeconcept to the crystallite-gel model for extrusion and spheronization. Pharm.Res. 15 509-510.

Ek, R., 1995. Characterisation of powder cellulose, microcrystalline cellulose and porous cellulose beads.Doctoral Thesis, University of Uppsala, Sweden.

El Saleh, F., Jumaa, M., Hassan, I. and Kleinebudde, P., 2000. Influence of cellulose typeon the properties of extruded pellets. II. Production and properties of pellets. S.T.P.Pharma Sciences 10 379-385.

Emerton, H.W., 1980. The preparation of pulp fibers for papermaking. In Handbook ofPaper Science, Vol 1, The Raw Materials and Processing of Papermaking, Rance, H.F.(Ed.), Elsevier Scientific Publishing Company, The Netherlands, pp. 139-164.

Faure, A., Grimsey, I., Rowe, R., York, P. and Cliff, M., 1999. Process control in a highshear mixer-granulator using wet mass consistency: The effect of formulation variables.J.Pharm.Sci. 88 191-195.

Fielden, K.E. and Newton, J.M., 1992. Extrusion and Extruders. In: Encyclopedia ofPharmaceutical Technology. Swarbrick J. and Boylan J.C., (Eds.), Vol. 5, Marcel Dekker, Inc.,New York, 395-442.

Fielden, K.E., Newton, J.M. and Rowe, R.C., 1992. The influence of lactose particle sizeon spheronization of extrudate processed by a ram extruder. Int.J.Pharm. 81 205-224.

Fielden, K.E., Newton, J.M., O’Brien, P. and Rowe, R.C., 1988. Thermal studies on theinteraction of water and microcrystalline cellulose. J.Pharm.Pharmacol. 40 674-678.

Fornés, V. and Chaussidon, J., 1978. An interpretation of the evolution with temperatureof the ν2 + ν3 combination band in water. J.Chem.Phys. 68 4667-4671.

Frake, P., Greenhalgh, D., Grierson, S.M., Hempenstall, J.M. and Rudd, D.R., 1997.Process control and end-point determination of a fluid bed granulation by application ofnear infra-red spectroscopy. Int.J.Pharm. 151 75-80.

Frank, H. and Wen, W-Y., 1957. Structural aspects of ion-solvent interaction in aqueoussolutions: A suggested picture of water structure. Discuss.Faraday Soc. 24 133-140.

Froix, M. and Goedde, A., 1976. The effect of temperature on the cellulose/waterinteraction from NMR relaxation times. Macromol. 9 428-430.

Froix, M. and Nelson, R., 1975. The interaction of water with cellulose from nuclearmagnetic resonance relaxation times, Macromol. 8 726-730.

Habib, Y., Abramowitz, R., Jerzewski, R., Jain, N. and Shreeram, A., 1999. Is silicifiedwet-granulated microcrystalline cellulose better than orginal wet-granulatedmicrocrystalline cellulose? Pharm.Dev.Technol. 4 431-437.

51

Hancock, B.C., York, P. and Rowe, R.C., 1992. Characterization of wet masses using amixer torque rheometer: 2. Mixing kinetics. Int.J.Pharm. 83 147-153.

Harrison, P.J., 1982. Extrusion of wet powder masses. Doctoral Thesis, University of London,UK.

Harrison, P.J., Newton, J.M. and Rowe, R.C., 1985. The characterization of wet powdermasses suitable for extrusion/spheronization. J.Pharm.Pharmacol. 37 686-691.

Hatakeyama, H., Hatakeyama, T. and Nakamura, K., 1983. Relationship betweenhydrogen bonding and water in cellulose. J.Appl.Polym.Sci. Appl.Polym.Symp. 37 979-991.

Hollenbeck, G., Peck, G. and Kildsig, D., 1978. Application of immersional calorimetryto investigation of solid-liquid interactions: Microcrystalline cellulose – water system.J.Pharm.Sci. 67 1599-1606.

Imanidis, G., 1986. Untersuchungen über die Agglomerierkinetik und die elektrischeLeistungsaufnahme beim Granulierprozess im Schnellmischer. Doctoral Thesis, University ofBasel, Switzerland.

Ishikiriyama, K. and Todoki, M., 1995. Pore size distribution measurements of silica gelsby means of differential scanning calorimetry. II. Thermoporosimetry. J.Colloid InterfaceSci. 171 103-111.

Iwamoto, M., Uozumi, J. and Nishinari, K., 1987. Preliminary investigation of the state of waterin foods by near infrared spectoscopy. Int. NIR/NIT conference, Budapest, Hungary, 3-12.

Jayme, G., 1944. Mikro-Quellungsmessungen an Zellstoffen. Der Papier-Fabrikant/Wochenblatt für Papierfabrikation 6 187-194.

Jeffrey, G., 1997. An introduction to hydrogen bonding. Oxford University Press, New York,303 pp.

Jørgensen, A., Rantanen, J., Karjalainen, M., Räsänen, E. and Yliruusi, J., 2001. CCD-Raman spectroscopy in monitoring pseudopolymorphic transitions during wet granulation. 4th

International Symposium on Solid oral dosage forms, May 13-15, Malmö, Sweden, 11.

Kleinebudde, P., 1994a. Shrinking and swelling properties of pellets containingmicrocrystalline cellulose and low substituted hydroxypropylcellulose: I. Shrinkingproperties. Int.J.Pharm. 109 209-219.

Kleinebudde, P., 1994b. Shrinking and swelling properties of pellets containingmicrocrystalline cellulose and low substituted hydroxypropylcellulose: II. Swellingproperties. Int.J.Pharm. 109 221-227.

Kleinebudde, P., Schröder, M., Schultz, P., Müller, B., Waaler, T. and Nymo, L., 1999.Importance of the fraction of microcrystalline cellulose and spheronization speed on theproperties of extruded pellets made from binary mixtures. Pharm.Dev.Technol. 4 397-404.

Kleinebudde, P., 1997. The crystallite-gel-model for microcrystalline cellulose in wet-granulation, extrusion and spheronization. Pharm.Res. 14 804-809.

52

Kleinebudde, P., Jumaa, M. and El Saleh, F., 2000. Influence of degree of polymerizationon behaviour of cellulose during homogenization and extrusion/spheronisation. AAPSPharmsci 2 (3) article 21 (http://www.pharmsci.org/)

Knight, P., 1993. Evaluation of binding agents for the preparation of spherical granules byextrusion/spheronisation. Doctoral Thesis, University of London, UK.

Kontny, M. and Zografi, G., 1995. Sorption of water by solids. In Physical characterization ofpharmaceutical solids. Brittain, H. (Ed.), Marcel Dekker Inc., New York, pp. 387-418.

Kristensen, H.G., Holm, P., Jægerskou, A. and Schæfer, T., 1984. Granulation in highspeed mixers. Part 4: Effect of liquid saturation on the agglomeration. Pharm.Ind. 46 763-767.

Kristensen, H.G. and Schæfer, T., 1987. Granulation. A review on pharmaceutical wet-granulation. Drug Dev.Ind.Pharm. 13 803-872.

Kristensen, J., Schæfer, T. and Kleinebudde P., 2000a. Direct pelletization in a rotaryprocessor controlled by torque measurements. I. Influence of process variables.Pharm.Dev.Technol. 5 247-256.

Kristensen, J., Schæfer, T. and Kleinebudde P., 2000b. Direct pelletization in a rotaryprocessor controlled by torque measurements. II. Effects of changes in the content ofmicrocrystalline cellulose. AAPS PharmSci 2(3) article 24 (http://www.pharmsci.org/)

Landín, M., Martínez-Pacheco, R., Gómez-Amoza, J.L., Souto, C., Concheiro, A. andRowe, R.C., 1993a. Effect of country of origin on the properties of microcrystallinecellulose. Int.J.Pharm. 91 123-131.

Landín, M., Martínez-Pacheco, R., Gómez-Amoza, J.L., Souto, C., Concheiro, A. andRowe, R.C., 1993b. Effect of batch variation and source of pulp on the properties ofmicrocrystalline cellulose. Int.J.Pharm. 91 133-141.

Landín, M., Rowe, R.C. and York, P., 1995. Characterization of wet powder masses witha mixer torque rheometer. 3. Nonlinear effects of shaft speed and sample weight.J.Pharm.Sci. 84 557-560.

Landín, M., York, P., Cliff, M.J., Rowe, R.C. and Wigmore, A.J., 1996. Scale-up of apharmaceutical granulation in fixed bowl mixer-granulators. Int.J.Pharm. 133 127-131.

Lane, R. and Buckton, G., 2000. The novel combination of dynamic vapour sorptiongravimetric analysis and near infra-red spectroscopy as a hyphenated technique.Int.J.Pharm. 207 49-56.

Leuenberger, H., Bier, H.-P. and Sucker, H., 1979. Theory of the granulation-liquidrequirement in the conventional granulation process. Pharm.Tech.Int. 3 61-68.

Leuenberger, H., Bier, H.-P. and Sucker, H., 1981. Determination of the liquidrequirement for a conventional granulation process. Ger.Chem.Eng. 4 13-18.

Li, J.-X., Zhou, Y., Wu xiao, Y., Odidi, I. and Odidi, A., 2000. Characterization of wetmasses of pharmaceutical powders by triaxial compression test. J.Pharm.Sci. 89 178-190.

53

Li, T.-Q., 1991. Interactions between water and cellulose fibers, Application of NMR techniques.Doctoral Thesis, Royal Institute of Technology, Stockholm, Sweden.

Li, T.-Q., Henriksson, O., Klasen, T. and Ödberg, L., 1992. Water diffusion in woodpulp cellulose fibers studied by means of the pulsed gradient spin-echo method. J.ColloidInterface Sci. 154 305-315.

Lindberg, N.-O., Leander, L., Wenngren, L., Helgesen, H. and Reenstierna, R., 1974.Granulation in a change can mixer. Acta Pharm. Suec. 11 603.

Luner, P., Majuru, S., Seyer, J. and Kemper, M., 2000. Quantifying crystalline formcomposition in binary powder mixtures using near-infrared reflectance spectroscopy.Pharm.Dev.Technol. 5 231-246.

Luukkonen, P., Schæfer, T., Mäkelä, K., Södergård, U., Hellén, L. and Yliruusi, J., 1999.Evaluation of the optimum water level for pelletisation using a mixer torque rheometer. AAPSPharmSci, 1148, AAPS Annual Meeting and Exposition, November 14-18, 1999, NewOrleans, USA.

Malamataris, S. and Kiortsis, S., 1997. Wettability parameters and deformationalbehaviour of powder-liquid mixes in the funicular agglomeration phase. Int.J.Pharm. 1549-17.

Maloney, T.C., 1999. Thermoporosimetry by isothermal step melting. ISWPC Pre-Symposium,Seoul, 245-253.

Maloney, T.C., 2000. On the pore structure and dewatering properties of the pulp fiber cell wall.Doctoral Thesis, Helsinki University of Technology, Finland.

Maloney, T.C., Paulapuro, H. and Stenius, P., 1998. Hydration and swelling of pulp fibersmeasured with differential scanning calorimetry. Nord.Pulp Pap.Res.J. 13 31-36.

Maloney, T.C., Li, T., Weise, U. and Paulapuro, H., 1997. Intra- and interfibre poreclosure in wet pressing. Appita J. 50 301-306.

Miwa, A., Yajima, T. and Itai, S., 2000. Prediction of suitable amount of water additionfor wet granulation. Int.J.Pharm. 195 81-92.

Miyake, Y., Shinoda, A., Uesugi, K., Furukawa, M. and Nasu, T., 1973. The influence ofamount of water on granulating efficiency and physical properties of spherical granulesprepared by extrusion-spheronization processing. Yakuzaigaku 33 167-171.

Morris, K., Stowell, J., Byrn, S., Placette A., Davis, T. and Peck G., 2000. Acceleratedfluid bed drying using NIR monitoring and phenomenological modeling. DrugDev.Ind.Pharm. 26 985-988.

Morris, K., Griesser, U., Eckhardt, C. and Stowell, J., 2001. Theoretical approaches tophysical transformations of active pharmaceutical ingredients during manufacturingprocesses. Advanced Drug Delivery Rewievs 48 91-114.

Mousseri, J., Steinberg, M.P., Nelson, A.I. and Wei, L.S., 1974. Bound water capacity ofcorn starch and its derivatives by NMR. J.Food Sci. 39 114-116.

54

Nakai, Y., Fukuoka, E., Nakajima, S. and Hasegawa, J., 1977. Crystallinity and physicalcharacteristics of microcrystalline cellulose. Chem.Pharm.Bull. 25 96-101.

Nakamura, K., Hatakeyama, T. and Hatakeyama, H., 1981. Studies on bound water ofcellulose by differential scanning calorimetry. J.Text.Inst. 72 607-613.

Nelson, R., 1977. The determination of moisture transitions in cellulosic materials usingdifferential scanning calorimetry. J.Appl.Polym.Sci, 21 645-654.

Newitt, D.M. and Conway-Jones, J.M., 1958. A contribution to the theory and practise ofgranulation. Trans.Inst.Chem.Eng. 36 422-442.

Newton, J.M., Chow, A.K. and Jeewa, K.B., 1992. The effect of excipient source onspherical granules made by extrusion/spheronisation. Pharm.Tech.Int. Oct, 52-58.

Nürnberg, E. and Wunderlich, J., 1999a. Manufacturing pellets by extrusion andspheronization. The importance of the π saturation value of water of lyoextrusion usingdifferent types of extruder. Part I. Pharm.Tech.Europe 11(2) 41-47.

Nürnberg, E. and Wunderlich, J., 1999b. Manufacturing pellets by extrusion andspheronization. The importance of the π saturation value of water of lyoextrusion usingdifferent types of extruder. Part II. Pharm.Tech.Europe 11(3) 30-34.

Oksanen, C. and Zografi, G., 1990. The relationship between the glass transitiontemperature and water vapor absorption by poly(vinylpyrrolidone). Pharm.Res. 7 654-657.

Osborne, B.G., Fearn, T. and Hindle, P.H., 1993. Practical NIR spectroscopy with applicationsin food and beverage analysis. Longman, Scientific and Technical, Harlow. 227 pp.

Ovenston, A. and Benbow, J.J., 1968. Effects of die geometry on the extrusion of clay-like material. Trans.Br.Ceram.Soc. 67 543-567.

Parker, M.D. and Rowe, R.C. 1991. Source variation in the wet massing (granulation) ofsome microcrystalline celluloses. Powder Technol. 65 273-281.

Parker, M.D., York, P. and Rowe, R.C., 1990. Binder-substrate interactions in wetgranulation. 1: The effect of binder characteristics. Int.J.Pharm. 64 207-216.

Parker, M.D., York, P. and Rowe, R.C., 1991. Binder-substrate interactions in wetgranulation. 2: The effect of binder molecular weight. Int.J.Pharm. 72 243-249.

Peemoeller, H. and Sharp, A.R., 1985. NMR study of cellulose-water systems: Waterproton spin-lattice relaxation in the rotating reference frame. Polymer 26 859-864.

Pizzi, A., Eaton, N.J. and Bariska, M., 1987a. Theoretical water sorption energies byconformational analysis. Part 1: Crystalline Cellulose 1. Wood Sci.Technol. 21 235-248.

Pizzi, A., Eaton, N.J. and Bariska, M., 1987b. Theoretical water sorption energies byconformational analysis. Part 2: Amorphous cellulose and the sorption isotherm. WoodSci.Technol. 21 317-327.

Podczeck, F., 1999a. Rheological studies of the physical properties of powders used incapsule filling. Part I. Pharm.Technol.Eur. 11(9) 16-24.

55

Podczeck, F., 1999b. Rheological studies of the physical properties of powders used incapsule filling. Part II. Pharm.Technol.Eur. 11(10) 34-42.

Podczeck, F. and Révész, P., 1993. Evaluation of the properties of microcrystalline andmicrofine cellulose powders. Int.J.Pharm. 91 183-193.

Podczeck, F. and Newton, J.M., 2000. Powder and capsule filling properties of lubricatedgranulated cellulose powder. Eur.J.Pharm.Biopharm. 50 373-377.

Raines, C.L., 1990. The extrusion of various formulations of microcrystalline cellulose. DoctoralThesis, University of London, UK.

Raines, C.L., Newton, J.M. and Rowe, R.C., 1989. Extrusion of microcrystalline celluloseformulations. In Proceedings of the Annual Meeting of the British Society of Rheology, Carter, R.E.(Ed.), Elsevier Applied Science, London, UK, pp. 248-257.

Rantanen, J., Antikainen, O., Mannermaa, J-P. and Yliruusi, J., 2000a. Use of near-infrared reflectance method for measurement of moisture content during granulation.Pharm.Dev.Technol. 5 209-217.

Rantanen, J., Jørgensen, A., Räsänen, E., Luukkonen, P., Airaksinen, S., Raiman, J.,Hänninen, K., Antikainen, O. and Yliruusi, J., 2001. An insight into fluidised bedgranulation process with in-line NIR moisture measurement. Submitted for publication.

Rantanen, J., Känsäkoski, M., Suhonen, J., Tenhunen J., Lehtonen, S., Rajalahti, T.,Mannermaa, J-P. and Yliruusi, J., 2000b. Next generation fluidized bed granulatorautomation, AAPS PharmSciTech 1(2), http://www.pharmscitech.com/

Rantanen, J., Lehtola, S., Rämet, P., Mannermaa, J-P. and Yliruusi, J., 1998. On-linemonitoring of moisture content in an instrumented fluidized bed granulator with a multi-channel NIR moisture sensor. Powder Technol. 99 163-170.

Rowe, R.C. and Sadeghnejad, G.R., 1987. The rheology of microcrystalline cellulosepowder/water mixes – measurement using a mixer torque rheometer. Int.J.Pharm. 38 227-229.

Rowe, R.C., 1995. The rheological properties of lactose/microcrystalline cellulose/watermixes: Measurement using mixer torque rheometry. Pharm.Sci. 1 547-549.

Rowe, R.C., 1996. Characterization of wet powder masses using a mixer torquerheometer. 4. Effect of blade orientation. Int.J.Pharm. 133 133-138.

Rumpf, H., 1958. Grundlagen und Methoden des Granulierens. 1.Teil: Begriffe,Anwendung und Eigenschaften der Granulate. Chem.Ing.Techn. 30 144-158.

Räsänen, E., Rantanen, J., Jørgensen, A., Karjalainen, M., Paakkari, T. and Yliruusi, J.,2001. Novel identification of pseudopolymorphic changes of theophylline during wetgranulation using near infrared spectroscopy. J.Pharm.Sci. 90 389-396.

Sastry, K.V.S. and Fuerstenau, D.W., 1973. Mechanisms of agglomerate growth in greenpelletization. Powder Technol. 7 97-105.

56

Scallan, A.M. and Carles, J.E., 1972. The correlation of water retention value with fibersaturation point. Svensk Papperstidn. 75 699-703.

Schmidt, C. and Kleinebudde, P., 1999. Influence of the granulation step on pelletsprepared by extrusion/spheronization. Chem.Pharm.Bull. 47 405-412.

Seyer, J., Luner, P. and Kemper, M., 2000. Application of diffuse reflectance near-infrared spectroscopy for determination of crystallinity. J.Pharm.Sci. 89 1305-1316.

Shah, R., Kabadi, M., Pope, D. and Augsburger, L., 1995. Physico-mechanicalcharacterization of the extrusion-spheronization process. Part II: Rheologicaldeterminants for successful extrusion and spheronization. Pharm.Res. 12 496-507.

Sherrington, P.J., 1968. The granulation of sand as an aid to understanding fertilizer granulation.Aggregation, manufacture of granules, pellets and tablets. Instn.Chem.Eng.Symp. CE201-CE215.

Sherwood, B. and Becker, J., 1998. A new class of high-functionality excipients: Silicifiedmicrocrystalline cellulose. Pharm.Technol. 22 78-88.

Souto, C., Alvarez, R., Duro, R., Concheiro, A., Gomez-Amoza, J. L. and Martínez-Pacheco, R., 1998. Versatility of mixer torque rheometry predictions in extrusion-spheronization.Proc. 2nd World Meeting APGI/APV, 447-448.

Staniforth, J., Baichwal, A., Hart, J. and Heng, P., 1988. Effect of addition of water onthe rheological and mechanical properties of microcrystalline celluloses. Int.J.Pharm. 41231-236.

Staniforth, J.N. and Chatrath, M., 1996. Towards a new class of high functionality tabletbinders, I: Quasi-hornification of microcrystalline cellulose and loss of functionality.Pharm.Res. 13 S208.

Stone, J.E., Scallan, A.M., and Abrahamson, B., 1968. Influence of beating on cell wallswelling and internal fibrillation. Svensk Papperstidn. 19 687-694.

Svedas, V., 2000. Transformation of the near-infrared bands of cellulose surfacehydroxyls under the influence of adsorbed water molecules. Applied Spectrocopy 54 420-425.

Taylor, L. and Langkilde, F., 2000. Evaluation of solid-state forms present in tablets byraman spectroscopy. J.Pharm.Sci. 89 1342-1353.

Thiel, P.A. and Madey, T.E., 1987. The interaction of water with solid surfaces:fundamental aspects. Surface Science Reports 7 211-385.

Thomson, W., 1871. On the equilibrium of vapour at a curved surface of liquid. Phil.Mag.42 448-452.

Tobyn, M.J., McCarthy, G., Staniforth, J. and Edge, S., 1998. Physicochemicalcomparison between microcrystalline cellulose and silicified microcrystalline cellulose.Int.J.Pharm. 169 183-194.

Tomer, G. and Newton, J.M., 1999a. A centrifuge technique for the evaluation of theextent of water movement in wet powder masses. Int.J.Pharm. 188 31-38.

57

Tomer, G. and Newton, J.M., 1999b. Water movement evaluation during extrusion ofwet powder masses by collecting extrudate fractions. Int.J.Pharm. 182 71-77.

Tomer, G., Newton, J.M. and Kinchesh, P., 1999a. Magnetic resonance imaging (MRI) asa method to investigate movement of water during the extrusion of pastes. Pharm.Res. 16666-671.

Tomer, G., Mantle, M.D., Gladden, L.F. and Newton, J.M., 1999b. Measuring waterdistribution in extrudates using magnetic resonance imaging. Int.J.Pharm. 189 19-29.

Tomer, G., Patel, H., Podczeck, F. and Newton, J.M., 2001. Measuring the waterretention capacities (MRC) of different microcrystalline cellulose grades. Eur.J.Pharm.Sci.12 321-325.

Travers, D.N., Rogerson, A.G. and Jones, T.M., 1975. A torque arm mixer for studyingwet massing. J.Pharm.Pharmacol. 27 Suppl. 3P.

Tuleu, C. and Chaumeil, J.C., 1998. Small-scale characterization of wet powder massessuitable for extrusion-spheronization. Drug Dev.Ind.Pharm. 24 423-429.

Urquhart, A., 1924. The mechanism of the adsorption of water by cotton. J.Text.Inst. 20T125-132.

Usteri, M. and Leuenberger, H., 1989. Agglomeration of binary mixtures in a high-speedmixer. Int.J.Pharm. 55 135-141.

Van Campen, I., Amidon, G.I. and Zografi, G., 1983. Moisture sorption kinetics forwater-soluble substances I: Theoretical considerations of heat transport control.J.Pharm.Sci. 72 1381-1388.

Watano, S., Takashima, H., Sato, Y., Miyanami, K. and Terashita, K., 1996. IR absorptioncharacteristics of an IR moisture sensor and mechanism of water transfer in fluidized bedgranulation. Advanced Powder Technol. 7 279-289.

Watano, S., Terashita, K. and Miyanami, K., 1990. Development and application ofinfrared moisture sensor to complex granulation. Bull.Univ.Osaka Pref. Series A 39 187-197.

Weise, U., 1997. Characterization and mechanism of changes in wood pulp fibers caused by waterremoval. Doctoral Thesis, Helsinki University of Technology, Finland.

Weise, U., 1998. Hornification - mechanisms and terminology. Paper and Timber 80 110-115.

Westermarck, S., Juppo, A.M., Kervinen, L. and Yliruusi, J., 1999. Microcrystallinecellulose and its microstructure in pharmaceutical processing. Eur.J.Pharm.Biopharm. 48199-206.

Vervaet, C., 1997. Influence of process and formulation parameters on the quality of pellets made byextrusion/spheronisation. Doctoral Thesis, Gent University, Belgium.

White, J., 1994. On-line moisture detection for a microwave vacuum dryer. Pharm.Res. 11728-732.

58

Wu, J.-S., Ho, H.-O. and Sheu M.-T., 2001. A statistical design to evaluate the influenceof manufacturing factors on the material properties and functionalities ofmicrocrystalline cellulose. Eur.J.Pharm.Sci. 12 417-425.

Yamauchi, T. and Murakami, K., 1991. Differential scanning calorimetry as an aid forinvestigating the wet state of pulp. J.Pulp Pap.Sci. 17 J223-226.

Young, J.H. and Nelson, G.L., 1967a. Research of hysteresis between sorption anddesorption isotherms of wheat. Trans.Am.Soc.Agric.Engs. 10 756-761.

Young, J.H. and Nelson, G.L., 1967b. Theory of hysteresis between sorption anddesorption isotherms in biological materials. Trans.Am.Soc.Agric.Engs. 10 260-263.

Zeronian, S.H., 1985. Intercrystalline swelling of cellulose. In Cellulose Chemistry and ItsApplications, Nevell T.P. and Zeronian S.H. (Eds), Ellis Horwood Limited, Great Britain,pp.138-158.

Zografi, G., Kontny, M.J., Yang, A.Y.S. and Brenner, G.S., 1984. Surface area and watervapor sorption of microcrystalline cellulose. Int.J.Pharm. 18 99-116.

Zografi, G. and Kontny, M., 1986. The interactions of water with cellulose- and starch-derived pharmaceutical excipients. Pharm.Res. 3 187-194.

Zografi, G., 1988. States of water associated with solids. Drug Dev.Ind.Pharm. 14 1905-1926.