Compaction of Powder Blends

Effect of pores, particles and percolation on tablet strength

Rijksuniversiteit Groningen

Compaction of Powder Blends Effect of pores, particles and percolation on tablet strength

Proefschrift ter verkrijging van het doctoraat in de Wiskunde en Natuurwetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op vrijdag 25 april 2003 om 14.15 uur door Bert van Veen geboren op 11 mei 1976 te Gersloot (Heerenveen)

Promotor Prof. Dr. H.W. Frijlink Co-Promotores Dr. G.K. Bolhuis Dr. Ir. K. van der Voort Maarschalk Beoordelingscommissie Prof. Dr. J. Ketolainen Prof. Dr. H. Vromans Prof. Ir. H. Wesselingh

Foar s Heit en Mem

The research presented in this thesis is part of the research school Groningen University Institute for Drug Exploration (GUIDE). Printing of this thesis was financially sponsored by: -N.V. Organon -GUIDE -Rijksuniversiteit Groningen

Printing: PrintPartners Ipskamp, Enschede, The Netherlands ISBN: 90-9016759-5

Contents

Chapter 1 Chapter 2

Introduction Tensile strength of tablets containing two materials with a different compaction behaviour Pore formation in tablets compressed from binary mixtures as a result of deformation and relaxation of particles The influence of particles of a minor component on the matrix strength of sodium chloride The mutual effects of initial particle size and addition of another compound on tablet strength A practical approach towards a fit to predict mechanical properties of compacts containing two components Conclusions and future research Summary Samenvatting Dankwoord

1 23

Chapter 3

37

Chapter 4

53

Chapter 5

69

Chapter 6

83

Chapter 7

103 107 111 117

Introduction

Chapter 1

Chapter 1 Introduction
1.1 General introduction

The tablet is still the most frequently administered dosage form for medical applications (1). A tablet has numerous advantages over other dosage forms, among which the convenient way of administration to patients and the stability of a drug substance in a dry dosage form (1). Furthermore, tablets are easy to handle, to store, to distribute and to manufacture in large quantities. A tablet is not only a carrier of the active component, the drug substance, but must also comply with a whole scale of demands. One of these demands is that the strength of a tablet should be sufficiently high to withstand the different handlings in the logistic chain from producer to patient. Basically, tablets are produced by the compaction of powders. The compaction of powders preceded only by mixing of the individual materials is called direct compaction. To make tablet production by direct compaction possible, it is necessary that the powders to be compacted fulfil a number of criteria (1). Therefore, various excipients specifically designed for direct compaction have been introduced during the last forty years (2). For an excipient to be useful as a filler-binder, it should meet the requirements as given in Table 1.1.
Table 1.1. Properties of an ideal filler-binder (direct compression excipient). 1. 2. 3. 4. 5. 6. 7. 8. The material should enhance the tablet strength. It should have a good compaction pressure-tablet strength profile. It should have a low lubricant sensitivity. It should have a good flowability. It should be compatible with all types of drug substances. It should not interfere with the bioavailability of drug substances. Mixtures with various drug substances should have a low tendency for segregation. It should not show any physical or chemical change on ageing and should be stable to air, moisture and heat. 9. It should have a high dilution capacity, which is defined as the amount of active ingredients that the filler-binder can successfully carry in the direct compaction technique. The dilution capacity is generally expressed in terms of percentage of non-compressible material or as optimum drug to filler-binder ratio. 10. It should be relatively cost effective.

Chapter 1

Introduction

However, not a single excipient fulfils all these requirements. It is general practise to use a combination of two (sometimes more) filler-binders in a tablet formulation in order to obtain a compact with acceptable properties. Among the various requirements, one of the most important functions of a filler-binder is the enhancement of the tablet strength. For this reason, most of the research on direct compaction focuses on the compaction behaviour and binding properties of existing and new filler-binders. Unfortunately, studies often investigate and report only the compaction behaviour of a filler-binder as a single component and hardly in combination with other filler-binders or other main components in the tablet formulation. Tablets produced by the pharmaceutical industry contain several ingredients. In addition to one or multiple filler-binders and drug substances, generally a disintegrant, a lubricant and other substances such as glidants are present in the tablet formulation. It should be realised that every component in a formulation may affect the compaction properties of the powder blend and may influence the tablet strength or other mechanical properties. Even low-concentrated additives in a tablet formulation can have enormous effects on the tablet strength. For example, the interactions between lubricants and filler-binders in relation to tablet strength are well described (e.g. 3, 4). Although a tablet commonly contains combinations of filler-binders and drug substances, only a limited number of studies have investigated the interactions between two main components during the compaction of a binary powder blend (5). So far, the complexity of all different binary mixtures made it impossible to establish a consistent theory to describe the compaction behaviour of a mixture of two materials and the resulting tablet strength. As physically blended binary mixtures are mixed powders containing single particles of two materials, this introduction will start by discussing the densification mechanism (section 1.2) and relaxation behaviour (section 1.3) of powders consisting of a single material. At the end of these sections, the few studies that have been published on binary mixtures are summarised. Section 1.4 describes the internal structure of a tablet, which is a result of the phenomena mentioned in sections 1.2 and 1.3. This final tablet structure is important for the different tablet properties. Section 1.5 reports the relations between the tablet structure and the mechanical properties of porous compacts and describes the theories developed on tablet strength. The structural changes in tablets compressed from a single material and from binary mixtures are described by the percolation theory in section 1.6. This section also gives a summary of reported tablet strength changes of tablets compressed from binary mixtures. Finally, the aim of this thesis is given in section 1.7.

Introduction

Chapter 1

1.2 1.2.1

Densification of a powder bed Densification mechanisms

The compaction of a tablet starts with applying pressure to a powder bed. The bulk volume of the powder bed decreases as an effect of a series of processes that happen sequentially or in parallel (6). The first stage of the consolidation process is characterised by particle rearrangement at low pressures. At a certain relative density, the particles are more or less in a fixed arrangement. Due to the resistance of a material against deformation (strain), the stress inside the particles increases. If the applied stress is released before the deformation reaches a specific critical value, the particles deform elastically, i.e. the deformation is reversible, and the particles inside the powder bed regain their original shapes. Until this critical value, the stress is linearly proportional to the deformation and is characterised by the elastic or Youngs modulus (E, Fig. 1.1a). The following stage can be divided into two main deformation mechanisms: Particles fragment into smaller units at a certain stress value (f, Fig. 1.1b). This stress is the fracture strength. These materials are referred to as brittle materials. After a critical stress (y, Fig. 1.1c), the particles yield and start to deform plastically. This critical stress is the yield strength of a material. Whereas elastic deformation is a reversible process, plastic deformation results in a permanent change of the particle shape. These materials are referred to as ductile or plastic materials. Material fracture will eventually occur at higher deformations.

f
Stress

Fracture strength Fracture strength

Yield strength

(c2) (c1)

E (a)

E (b)

E (c) Strain

Figure 1.1. Macroscopic stress-strain relations: a. reversible elastic deformation; b. brittle behaviour; c. ductile behaviour (c1. normal plastic flow; c2. strain-hardening).

Chapter 1

Introduction

Whether particles behave ductile or brittle under compression depends on the material and its physical condition. Roberts and Rowe demonstrated (7,8) that materials possess a critical particle diameter at which the densification mechanism turns from brittle to ductile when the particle size decreases (Fig. 1.2). The stress necessary to cause particle fracture increases when the particle size decreases, whereas the stress causing plastic deformation of a material is independent of the particle size. When the fracture stress reaches the level of the yield strength, particles with diameters lower than the critical diameter will yield instead of fracturing. Many materials used as a filler-binder in the pharmaceutical industry can be classified into groups of materials that only have a brittle or ductile deformation behaviour due to their extremely low (e.g. inorganic compounds) or high (e.g. celluloses and starches) critical particle diameters (9). The critical particle diameters of other materials are close to the particle sizes of powders used in practice, such as -lactose monohydrate.

Stress Yield strength

Fracture strength dcrit Particle diameter

Fig. 1.2. Schematic representation depicting the effect of particle size on yield strength or fracture strength.

1.2.2

Characterisation of powder densification

The determination of the densification behaviour of a powder is usually based on the measurement of the relative density of the powder bed as a function of the pressure necessary to obtain that relative density. Table 1.2 shows an overview of relations used to describe the effect of the pressure on the degree of densification.

Introduction

Chapter 1

Table 1.2. Powder densification-compaction pressure (P) relations used to characterise the densification mechanism of a powder bed Name Heckel equation Equation Parameters r is the relative density of the powder bed. K and A are constants. Vi and Vp are the initial volume and volume under pressure of the powder, respectively. a and b are constants.

ln

1 = KP + A 1 r = abP 1 + bP

Kawakita equation

Vi V p Vp

Cooper and Eaton equation Walker and Balshin equation

Vi V p

K1 K 2 Vt is the true powder volume. = C1 exp + C 2 exp C1, C2, K1 and K2 are Vi Vt P P constants.
Vr is the relative density of the powder. C3 and K3 are constants.

Vr = C 3 K 3 log P

The equation most commonly employed in pharmaceutical compaction research is the Heckel-equation (10, 11). This equation assumes a linear section of the relationship between the logarithm of the relative density under pressure and the applied pressures (Table 1.2). Slope K is related to the yield strength (y, Fig 1.1c) of a material by:

y =

1 3K

(1)

The reciprocal of the slope K has been found to reflect the yield pressure Py of a material (12). This means that the yield pressure Py of a material is equal to 3y. The value A of the intercept with the y-axis has been related to the relative density of the fixed packing after particle rearrangement (10,11). As Heckel attributed the linear section to plastic deformation, equation 1 is in principle only appropriate to examine the densification behaviour of plastic materials. Nevertheless, Roberts et al. (13) modified the yield pressure Py to the deformation stress d, which is a plastic deformation stress, a fracture deformation stress or a combination of both, in order to calculate the fracture stress of particles with diameters larger than the critical diameter. Furthermore, particle size (e.g. 14, 15), compaction speed (e.g. 15, 16) and moisture content (e.g. 17, 18) influence the slope of the linear section of equation 1 and consequently the value of the yield pressure.

Chapter 1

Introduction

1.2.3

Densification behaviour of binary mixtures

The densification behaviour of a binary mixture is usually examined by the change in yield pressure as a function of the mixture composition. Binary mixtures can be divided into three categories with respect to the main densification behaviours of the single materials: ductile-ductile, brittle-brittle and ductile-brittle. The different changes in yield pressure caused by alterations in the mixture composition will be evaluated as schematically depicted in Fig. 1.3. The relationships between the yield pressure and the mixture composition were divided into groups based on three different changes: a linear interpolation of the individual yield pressures and positive or negative deviations of this linear relationship.

Py

+ 0 -

Mixture composition (v/ v)

Fig. 1.3. Schematically depicted positive (+) and negative (-) trends in yield pressure of binary mixtures compared to linear interpolation (0).

For mixtures consisting of two plastically deforming materials, the most frequently reported relationship between the yield pressure and the mixture composition is a linear one (19, 20). The explanation given in the literature for this observation is that the particles of the materials will not obstruct or enhance the yielding of the mixture. The yield pressure of a powder mixture is the sum of the relative contributions of both yield pressures. The yield pressures of binary mixtures containing two brittle components show no direct dependency on the mixture composition (19). A positive deviation from linear interpolation can be explained by the fact that a network of the more rigid particles in the powder mixture tends to support the applied load on the powder bed (19, 21). The opinion given by Ilkka and Paronen (19) is that the more fragile particles will be protected from fracture by this more rigid network. These explanations are

Introduction

Chapter 1

thought to be questionable. The fragile particles will weaken the network of the more rigid particles. Most of the fragile particles will already fracture at low stress values at the initial phase of the compression process. The most frequently investigated binary mixtures are the blends of ductile particles and particles that behave brittle under compression. For these blends, two main results for the change in yield pressure have been reported: a linear relationship (20, 22), and slight negative deviations from this linear relationship (19, 23, 24). To explain the lower yield pressures than expected from linear interpolation, Ilkka and Paronen (19) suggested that the plastically flowing material at higher concentrations occupies the pore spaces around the brittle particles preventing the stress to reach the fracture stress. Additionally, Sheikh-Salem and Fell (23) stated that the early fragmentation of high amounts of the brittle component might slow down the yielding of the plastically deforming material.

1.3 1.3.1

Tablet relaxation Characterisation of tablet relaxation

During and after the removal of the applied load, a compressed powder bed generally shows an elastic recovery of its volume as an effect of stress relaxation. In case of a permanent deformation, the volume expansion will be small, whereas large tablet relaxations are caused by an important elastic component (6). The driving force for tablet relaxation has been suggested to be the elastic energy stored in the particles during densification (25, 26). On the other hand, the bonding between particles is regarded as the counteracting force against tablet relaxation. Tablet relaxation is commonly expressed by a change in tablet height, volume or porosity (25, 27). During densification, some compacts may even show an increase in density above their own material density at high compaction pressures. The total tablet relaxation can be a combination of a fully reversible decrease of the true density to the material density and an increase in tablet porosity. The tablet expansion expressed as porosity expansion was found to be independent of the applied load (25). Similarly as has been described for the densification of a powder bed (section 1.2.1), the relaxation propensity of a tablet is influenced by the compaction speed (15, 28), moisture content (29, 18) and particle size (14, 15). Malamataris et al. (30) defined the plasto-elasticity of materials as the ratio between the stress relaxation (change in tablet height under maximum load for a

Chapter 1

Introduction

certain period of time; SR) and the elastic recovery (change in tablet height after ejection from the die; ER). This SR/ER ratio describes the ability of a material to deform plastically under load and recover elastically after the application of the load. 1.3.2 Relaxation behaviour of binary mixtures

Mixtures containing a brittle and a ductile material are interesting tools to examine the effect of relaxation behaviours of the individual components on the total tablet relaxation. Most authors (31, 32, 33) reported no direct relationship between elastic recovery of tablets consisting of blends and their compositions. Tablets compressed from mixtures show capping or lamination when the tablets contain less than about 15-25% of the ductile material and/or when the ratio between the stress relaxation and elastic recovery (SR/ER) is higher than 9 (30, 34). In contrast, Sheikh-Salem et al. (31) have found a lower critical value of 1.5 for the SR/ER ratio of tablets that showed a significant capping behaviour. A possible explanation of the discrepancy between these two values (1.5 and 9) given by the authors (31) is the correction of the punch deformations, which gives a better estimation of the powder bed height under pressure. However, in all three studies (30, 31, 34) the tensile strength of tablets compressed from paracetamol-microcrystalline cellulose blends could be related to the SR/ER-ratio.

1.4

Tablet structure

The final tablet structure is the combined result of the consolidation of the powder bed and relaxation of the tablet. As both processes are influenced by the different factors mentioned in section 1.2.1 and 1.3.1, the final tablet structure also depends on these factors. In all situations, the internal tablet structure can be seen as a combination of particles and voids. But, the axial compaction of powders leads to anisotropy and heterogeneous tablet structures (35). The predominant bonding mechanisms between particles in compacts compressed from pharmaceutical materials are intermolecular forces, i.e. Van der Waals forces, electrostatic forces and hydrogen bonding (36), and solid bridge formation in case of salts such as sodium chloride, potassium chloride and sodium bicarbonate (37). Different approaches and techniques have been developed to describe and measure quantitatively the surface energy and bonding capacity of materials, e.g. the solubility parameter, which equals the root of the cohesive energy density (38), atomic force

Introduction

Chapter 1

measurements (39), inverse gas chromatography (40) or tablet strength measurements in liquids with different dielectric constants (41, 42). With the help of the solubility parameter, theoretical estimations of an adhesive interaction parameter can be made between different materials in powder blends (43, 44). The particles inside the tablet structure directly influence the size of voids. Mercury intrusion porosimetry, a commonly used technique to measure the pore size distribution, is based on the pressures necessary to enforce mercury into pore openings. According to Bockstiegels observation for compacts of iron particles (45, 46), tablets compacted from ductile materials show a decrease of the larger pores at increasing compaction pressures, while the smaller pores do not change in size and number (47, 48, 49). The fragmentation of particles of a brittle material results in an increase of smaller particles and consequently a different change in the pore size distribution as an effect of compression. The disappearance of larger pores is accompanied by an increase in the amount of small pores (48, 49, 50). In tablets compressed from ductile materials, the pore size will be dependent on the initial particle size. Brittle fracture will to a large extent eliminate the effects of the initial particle size.

1.5 1.5.1

Mechanical properties of compacts Porosity-mechanical properties relationships

Mechanical properties of compacts such as tensile strength, elastic modulus and indentation hardness are known to be largely dependent on the compact structure. Mechanical properties commonly decrease with an increase of the tablets porosity. The change in mechanical property as a function of the porosity is normally fitted by an exponential relationship (51, 52). The extrapolated value at zero porosity of this function is normally regarded as an indication of the material property. However, this interpretation has been questioned in several studies; different authors showed that a smaller initial particle size leads to an increase of the mechanical property (53, 54). However, it is interesting to note that direct relationships have been demonstrated between the calculated cohesive energy densities of several materials and their elastic moduli (55), tensile strengths (56) and the critical stress intensity factors (56) (see sec. 1.5.3). The tablet relaxation after compaction results in an increase of the tablet porosity and consequently into a decrease of the tensile strength. Remarkably, some crystalline

Chapter 1

Introduction

materials, such as sodium chloride and potassium chloride, have been found to give an increase in tablet strength after compaction (57, 58). The authors postulated that the stress relaxation caused by recrystallisation creates extra bonding forces between particles. 1.5.2 Bond summation mechanism

Two main theories have been developed to describe the important factors controlling the strength of a porous compact. One of these theories is the bond summation mechanism. The tablet strength is the sum of interparticle bonding forces between surfaces with spherical curvatures in the diametrical failure plane of the compact (59, 60). The surface area taking part in the interparticle attraction is small and difficult to measure. Therefore, secondary definitions of bonding surface areas of powders and tablets are used to explain changes in tablet strength (37, 61). The specific surface area of tablets compressed from crystalline lactoses has been found to be directly related to the tablet strength (49). A similar relationship has been reported for tablets compressed from binary mixtures of different types of crystalline lactose (62). However, this relationship could not be established for other materials (37, 61). Recently, views are more focused on fracture mechanism theory. 1.5.3 Fracture mechanism

The fracture mechanism theory is based on the assumption that pores in the tablet structure act as stress concentrators. Fracture of a porous material occurs when the stress of the pores at the bonding sites of the particles is higher than the cohesive energy of the material and high enough to initiate fracture. To continue tablet fracture, the elastic energy released must be higher than the energy necessary to form new pore openings. Inglis (63) was the first author to postulate that pores or cracks perpendicular to the applied load concentrate stress at their tips and consequently decrease the fracture strength of a material (Fig. 1.4). The ratio between the material strength with or without a crack for elastic-brittle fracture is given by:

material l = 1+ 2 measured

(2)

where material is the cohesive strength of the material, measured the measured compact strength, l the length of the pore and the diameter of the curvature of the pore tip.

10

Introduction

Chapter 1

measured

measured

material

l measured

material d

measured

Fig. 1.4. Inglis-Griffith crack theory

Griffith (64) further developed equation 2 into a universal theory based on thermodynamical considerations. He argued that two energies have to be taken into account when a crack propagates: a release of elastic energy that is stored in the material and an increase of surface energy. The combination of these two energies resulted into:

measured =

2 E s a

(3)

with E is the elastic modulus of the body, s the surface energy and a is the half of pore length l. The stress around a crack in a material is reflected by the critical stress intensity factor Kic (= (2Es)) and is a measure of the stress necessary to start and develop a failure surface (65). The theory of elastic-brittle fracture can also be applied to tablet fracture. Stanley (66) states that the fracture of porous compacts behaves in a brittle fashion, i.e. no plastic deformation occurs before fracture starts. As defined by Griffiths equation, larger cracks or pores give a higher decrease of the tablet strength. In literature, there are indications that the tablet strength is related to the median pore diameter, which is primarily affected by the larger pores (50, 67, 68).

11

Chapter 1

Introduction

1.6 1.6.1

Compacts consisting of binary mixtures Percolation theory

During the last two decades, percolation theory has been used to explain macroscopic changes in compact structures. Percolation theory, developed since the 1970s and introduced by Stauffer (69) and Leuenberger (70, 71) in pharmaceutical research, regards a compact structure as a complex of sites. The sites can be occupied randomly by particles or pores (site-percolation). A group of sites consisting of particles is considered to belong to the same cluster when bonds are formed between the neighbouring particles (bond-percolation). The percolation theory describes phenomena whereby tablet properties divert, vanish or start to appear at a certain tablet composition. Particles of a material transform from fractional clusters to a continuous network (matrix), or vice versa, at this critical tablet composition, which is defined as the percolation threshold. Percolation theory can be used to describe structure changes in compacts compressed from single materials and from blends. Even a compact consisting of a single material has different percolation thresholds. The compact not only contains particles, but also pores between these particles and can therefore be regarded as a composite structure. At very low relative densities, particles of a material only form a loose powder. A compact with finite mechanical properties is first established when the relative density of the powder compact becomes larger than the relative tapped density (1). This relative tapped density is regarded as the percolation threshold of the solid matrix formed by particles. Above a certain relative density, the pores between the particles will no longer form a continuous network (2) which is the percolation threshold of air (72). According to the percolation theory, a property Y of a tablet changes in the vicinity of the percolation threshold of the material following a power law:

Y = Ymax ( r c )

(4)

where Ymax is the maximum value, r the relative density of the compact, c the critical relative density (the percolation threshold) and T the critical exponent. Practical cases already have shown that equation 4 is also valid for a broader range of relative densities higher than the percolation threshold (73). The critical exponent T reflects the rate of change in property Y towards the percolation threshold. The value of the critical exponent T depends on the different types of particle packaging (70, 71). In these studies, the relative tap density functions as the percolation threshold c of the solid particles in tablets consisting of a single material. The relative density, which

12

Introduction

Chapter 1

equals the volume fraction, was used as an indirect measure to explore the relations between the relative amount of interparticle bonds and the mechanical properties. In further research, the normalised solid fraction N was defined as a ratio between the number of bonds formed and the maximum number of bonds possible (72, 74), resulting in:

N=

r c 1 c

(5)

The normalised solid fraction N is a functional parameter describing the threedimensional bonding between particles in porous compacts. Assuming T to have a value of 1, the combination of equations 4 and 5 gives a general relationship for the change in property Y:

Y=

Y max ( r c ) 1 c

(6)

This equation is usually not obeyed: the normalised solid fraction reflects the dependence of property Y on three-dimensional bonding, and the change of property Y should be described by:

r c Y = Ymax 1 c

(7)

For describing the particle structure in compacts compressed from binary mixtures, site percolation is an important concept. At different concentrations of materials A and B, different percolation thresholds can be expected: At concentrations of A lower than the percolation threshold of material A, clusters of particles A are dispersed in a matrix of B. At concentrations of A between the percolation thresholds of both materials, both materials form an individual matrix. At concentrations of material A higher than the percolation threshold of material B, particles of material A only form an infinite matrix combined with clusters of particles B.

13

Chapter 1

Introduction

Actually, tablets compressed from binary mixtures contain a third component: air. The percolation threshold of air in tablets compressed from single materials and binary mixtures is usually found at a volume fraction of about 0.10 (72, 75). The percolation thresholds of materials in tablets compressed from binary mixtures depend on the differences in particle packaging, particle size, particle size distribution, shape of the particles and yield pressure between both materials (70, 76, 77). 1.6.2 Tablet strength of binary mixtures

Studies that investigated the tablet strength of compacts compressed from mixtures of two materials have reported a series of contradictory results. As in section 1.2.3, the mixtures were distinguished by the combination of the individual densification behaviour of the materials in the blends. The results of these studies were examined correspondingly as the change in yield pressure of powder blends (Fig. 3). For mixtures of which all particles deform plastically under densification, very different changes of the tablet strengths versus the mixture composition are observed. Some show a linear relationships (78, 79, 80) and others positive (81, 82) or negative (83, 84) deviations from this relation. Mattson and Nystrm (82, 85) researched the positive effect of a soft binder on the tablet strength of a ductile material. They found that the ability of a soft binder to fill voids between the compound particles plays a decisive role in the increase of the tablet strength. Most binary mixtures containing materials that have brittle densification behaviours show a similar relationship between the tablet strength and the relative proportions of the two components. The tablet strengths of these mixtures are mostly found to be linearly related to their compositions (62, 78, 86). The strength of a tablet consisting of a blend can be calculated from the tablet strengths of the pure materials. In contrast to tablets compressed from brittle-brittle mixtures, compacts consisting of a brittle and a ductile material rarely show a linear trend of the tablet strength as a result of composition changes. The strength of compacts produced from these mixtures is found to be higher (24, 77, 84, 87) or lower (23, 86, 83, 79) than the linear extrapolation of the individual material strengths. However, a reason for these, sometimes large, deviations in observed tablet strengths has not been presented. Different authors have postulated that a higher tablet strength can be caused by a higher bonding strength between both materials (42, 79), an increase of the number of interparticle bonding sites in the compact (82, 85), a higher densification of the mixture (82, 83, 85) or a combination of these phenomena.

14

Introduction

Chapter 1

1.7

Aim of this thesis

As described in section 1.6, many studies have been conducted to elucidate the change in tablet strength of compacts consisting of mixtures. However, it is impossible to draw straightforward conclusions because of the different experimental set-ups and results. Compaction pressure, particle size, compaction speed, moisture content and other factors that can influence the compact strength differ between these studies. As explained in sections 1.5 and 1.6, the final tablet strength, more than these factors, is directly related to the tablet structure. The change in tablet strength of mixtures should be based on the volume fractions of both materials, rather than the experimentally more convenient mass fractions. Proceeding on this line of thought, an important step to describe the change in strength of tablets compressed from blends as an effect of alterations in mixture composition is to choose a constant porosity instead of a constant compaction pressure. This is a different approach than usually found in other studies that have focused on this subject. To investigate the mechanisms that influence the densification behaviour, relaxation behaviour and final tablet strength of mixtures, the chapters in this thesis are based on studies in which only two model materials, sodium chloride and pregelatinised starch, will be used. The materials are selected on the following criteria: both materials are known to behave ductile during compaction, both materials differ in densification and relaxation behaviour, the chemical-physical properties of the single materials are well described and the materials are available as pure materials. This thesis investigates the changes in tablet structure of blends composed of the two model components. It aims to establish relations between the compaction behaviour of blends and the tablet strength. At the end, a model is presented to describe a relative particle interaction between both materials and to predict properties of tablets compressed from mixtures from the individual material properties.

15

Chapter 1

Introduction

1.8
1 2 3

References
Jivraj, M., Martini, L.G., Thompson, C.M., An overview of the different excipients useful for the direct compression of tablets. Pharm. Sci. & Techn. today, 3(2) (2000) 58-63. Armstrong, N.A., Selection of excipients for direct compression tablet formulations. Pharm. Techn. Europe, 9(8) (1997) 24-30. De Boer, A.H., Bolhuis, G.K., Lerk, C.F., Bonding characteristics by scanning electron microscopy of powders mixed with Magnesium Stearate. Powder Techn., 20 (1978) 7582. Zuurman, K., Van der Voort Maarschalk, K., Bolhuis, G.K., Effect of magnesium stearate on bonding and porosity expansion of tablets produced from materials with different densification behaviours. Int. J. Pharm., 179 (1999) 107-115. Fell, J.T., Compaction properties of binary mixtures. In: Alderborn, G., Nystrm, C. (Eds.), Pharmaceutical Powder Compaction Technology. Marcel Dekker, New York (1996) 501-515. Duberg, M., Nystrm, C., Studies on direct compression of tablets: XVII. Porositypressure curves for the characterization of volume reduction mechanisms in powder compression. Powder Techn., 46 (1986) 67-75. Roberts, R.J., Rowe, R.C., The compaction of pharmaceutical and other model materials a pragmatic approach. Chem. Eng. Sci., 42(4) (1987) 903-911. Roberts, R.J., Rowe, R.C., Brittle/ductile behaviour in pharmaceutical materials used in tabletting. Int. J. Pharm., 36 (1987) 205-209. Rowe, R.C., Roberts, R.J., Mechanical Properties, V. Effect of particle size-brittle/ductile transitions. In: Alderborn, G., Nystrm, C. (Eds.), Pharmaceutical Powder Compaction Technology. Marcel Dekker, New York (1996) 315-317. Heckel, R.W., Density-pressure relationships in powder compaction. Trans. Met. Soc. AIME., 221 (1961) 671-675. Heckel, R.W., An analysis of powder compaction phenomena. Trans. Met. Soc. AIME., 221 (1961) 1001-1008. Hersey, J.A., Rees, J., Deformation of particles during briquetting. Nature Phys. Sci., 230 (1971) 96. Roberts, R.J., Rowe, R.C., Kendall, K., Brittle-ductile transitions in die compaction of sodium chloride. Chem. Eng. Sci., 44(8) (1989) 1647-1651. Katikaneni, P.R., Upadrashta, S.M., Rowlings, C.E., Neau, S.H., Hileman, G.A., Consolidation of ethylcellulose: effect of particle size, press speed and lubricants. Int. J. Pharm., 117 (1995) 13-21. Nokhodchi, A., Rubinstein, M.H., An overview of the effects of material and process variables on the compaction and compression properties of hydroxypropyl methylcellulose and ethylcellulose. S.T.P. Pharma Sci., 11(3) (2001) 195-202. Roberts, R.J., Rowe, R.C., The effect of punch velocity on the compaction of a variety of materials. J. Pharm. Pharmacol., 37 (1985) 377-384.
16

7 8 9

10 11 12 13 14

15

16

Introduction 17

Chapter 1

Rees, J.E., Tsardaka, K.D., Some effects of moisture on the viscoelastic behaviour of modified starch during powder compaction. Eur. J Pharm. Biopharm., 40(4) (1994) 193197. Van der Voort Maarschalk, K., Vromans, H., Groenendijk, W., Bolhuis, G.K., Lerk, C.F., Effect of water on deformation and bonding of pregelatinized starch compacts. Eur. J. Pharm. Biopharm., 44 (1997) 253-260. Ilkka, J., Paronen, P., Prediction of the compression behaviour of powder mixtures by the Heckel equation, Int. J. Pharm., 94 (1993) 181-187. Duberg, M., Nystrm, C., Studies on direct compression of tablets: XII. The consolidation and bonding properties of some pharmaceutical compounds and their mixtures with Avicel 105. Int. J. Pharm. Techn. & Product Mfr., 6(2) (1985) 17-25. Bouvard, D., Densification behaviour of mixtures of hard and soft powders under pressure. Powder Techn., 111 (2000) 231-239. Humbert-Droz, P., Mordier, D., Doelker, E., Densification behaviour of powder mixtures. Acta Pharm. Techn., 29(2) (1983) 69-73. Sheikh-Salem, M., Fell, J.T., Compaction characteristics of mixtures of materials with dissimilar compaction mechanisms. Int. J. Pharm. Techn. & Product Mfr., 2(1) (1981) 19-22. Garr, J.S.M., Rubinstein, M.H., The effect of rate of force application on the properties of microcrystalline cellulose and dibasic calcium phosphate mixtures. Int. J. Pharm., 73 (1991) 75-80. Van der Voort Maarschalk, K., Zuurman, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., Porosity expansion of tablets as a result of bonding and deformation of particulate solids. Int. J. Pharm., 140 (1996) 185-193. Van der Voort Maarschalk. K., Zuurman, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., Stress relaxation of compacts produced from viscoelastic materials. Int. J. Pharm., 151 (1997) 27-34. Doelker, E., Comparative compaction properties of various microcrystalline cellulose types and generic products. Drug Dev. Ind. Pharm., 19 (1993) 2399-2471. Van der Voort Maarschalk, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., The effect of viscoelasticity and tabletting speed on consolidation and relaxation of a viscoelastic material. Eur. J. Pharm. Biopharm., 42 (1996) 49-55. Steendam, R., Frijlink, H.W., Lerk, C.F., Plasticisation of amylodextrin by moisture. Consequences for compaction behaviour and tablet properties. Eur. J. Pharm. Sci., 14 (2001) 245-254. Malamataris, S., Bin Baie, S., Pilpel, N., Plasto-elasticity and tableting of paracetamol, Avicel and other powders. J. Pharm. Pharmacol., 36 (1984) 616-617. Sheikh-Salem, M., Fell, J.T., Alkaysi, H.N., Mushin, N.A., Stress Relaxation of Mixtures of Pharmaceutical Powders. Acta Pharm. Techn., 30(4) (1984) 312-316. Nokhodchi, A., Rubinstein, M.H., The compaction properties of mixtures of ibuprofen and hydroxypropylmethyl cellulose. S.T.P. Pharma Sci., 8(6) (1998) 343-348.

18

19 20

21 22 23

24

25

26

27 28

29

30 31 32

17

Chapter 1 33 34

Introduction

Picker, K.M., The use of carrageenan in mixture with microcrystalline cellulose and its functionality for making tablets. Eur. J. Pharm. Biopharm., 48 (1999) 27-36. Bangudu, A.B., Pilpel, N., Effects of composition, moisture and stearic acid on the plastoelasticity and tableting of paracetamol-microcrystalline cellulose mixtures. J. Pharm. Pharmacol., 37 (1985) 289-293. Train, D., An investigation into the compaction of powders. J. Pharm. Pharmacol., 8 (1956) 745-760. Ferrari, F., Bertoni, M., Bonferoni, M.C., Rossi, S., Caramella, C., Nystrm, C., Investigation on bonding and disintegration properties of pharmaceutical materials. Int. J. Pharm., 136 (1996) 71-79. Adolfsson, , Gustafsson, C., Nystrm, C., Use of tablet tensile strength adjusted for surface area and mean interparticulate distance to evaluate dominating bonding mechanism. Drug Dev. Ind. Pharm., 25(6) (1999) 753-764. Rowe, R.C., Binder-Substrate Interactions in Tablets: A Theoretical approach based on solubility parameters. Acta Pharm. Techn., 34(3) (1988) 144-146. Sindel, U., Zimmermann, I., Measurement of interaction forces between individual powder particles using an atomic force microscope. Powder Techn., 117 (2001) 247-254. Levoguer, C., Butler, D., Thielmann, F., Williams, D., Characterizing the physicochemical properties of pharmaceutical materials. Pharm. Techn. Europe, 12 (11) (2000) 36-42. Karehill, P.G., Nystrm, C., Studies on direct compression of tablets: XXI. Investigation of bonding mechanisms of some directly compressed materials by strength characterization in media with different dielectric constants (relative permittivity). Int. J. Pharm., 62 (1990) 251-260. Adolfsson, , Caramella, C., Nystrm, C., The effect of milling and addition of dry binder on the interparticulate bonding mechanisms in sodium chloride tablets. Int. J. Pharm., 160 (1998) 187-195. Barra, J., Lescure, F., Falson-Rieg, F., Doelker, E., Can the organization of a binary mix be predicted from the surface energy, cohesion parameter and particle size of its components? Pharm. Res., 15(11) (1998) 727-736. Rowe, R.C., Interactions in the ternary powder system microcrystalline cellulose, magnesium stearate and colloidal silica - a solubility parameter approach. Int. J. Pharm., 45 (1988) 259-261. Bockstiegel, G., Relations between pore structure and densification mechanism in the compacting of iron powders I. Compacting properties in relation to the pore structure inside and in between powder particles. Int. J. Powder Metall., 2(4) (1966) 13-26. Bockstiegel, G., Relations between pore structure and densification mechanism in the compacting of iron powders. Int. J. Powder Metall., 3(1) (1967) 29-37. Westermarck, S., Juppo, A.M., Kervinen, L., Yliruusi, J., Pore structure and surface area of mannitol powder, granules and tablets determined with mercury porosimetry and nitrogen adsorption. Eur. J. Pharm. Biopharm., 46 (1998) 61-68.

35 36

37

38 39 40

41

42

43

44

45

46 47

18

Introduction 48 49

Chapter 1

Juppo, A.M., Porosity parameters of lactose, glucose and mannitol tablets obtained by mercury porosimetry. Int. J. Pharm., 129 (1996) 1-12. Vromans, H., de Boer, A.H., Bolhuis, G.K., Lerk, C.F., Kussendrager, K.D., Bosch, H., Studies on tableting properties of lactose. Part. 2. Consolidation and compaction of different types of crystalline lactose. Pharm. Weekbl. (Sci.) 7 (1985) 186-193. Juppo, A.M., Relationship between breaking force and pore structure of lactose, glucose and mannitol tablets. Int. J. Pharm., 127 (1996) 95-102. Duckworth, W.H., Discussion of Ryshkewitch paper by Winston Duckworth. J. Am. Ceram. Soc., 36 (1953) 68. Spriggs, R.M., Expression for effect of porosity on elastic modulus of polycrystalline refractory materials, particulary aluminium oxide, J. Am. Ceram. Soc., 44 (1961) 628629. Sun, C., Grant, D.J.W., Effects of initial particle size on the tableting properties of Llysine monohydrochloride dihydrate powder. Int. J. Pharm., 215 (2001) 221-228. Bin Baie, S., Newton, J.M., Podczeck, F., The characterization of the mechanical properties of pharmaceutical materials. Eur. J. Pharm. Biopharm., 43(2) (1996) 138-141. Roberts, R.J., Rowe, R.C., The relationship between Young's modulus of elasticity of organic solids and their molecular structure. Powder Techn., 65 (1991) 139-146. Roberts, R.J., Rowe, R.C., York, P., The relationship between the fracture properties, tensile strength and critical stress intensity factor of organic solids and their molecular structure. Int. J. Pharm., 125 (1995) 157-162. El Gindy, N.A., Samaha, M.W., Tensile strength of some pharmaceutical compacts and their relation to surface free energy. Int. J. Pharm., 13 (1983) 35-46. Eriksson, M., Alderborn, G., Mechanisms for post-compaction changes in tensile strength of sodium chloride compacts prepared from particles of different dimensions. Int. J. Pharm., 109 (1994) 59-72. Hiestand, E.N., Dispersion Forces and Plastic Deformation in Tablet Bond. J. Pharm. Sci., 74(4) (1985) 768-770. Rumpf, H., Zur theorie der zugfestigkeit von agglomeraten bei krafbertragung an kontaktpunkten. Chemie-Ing.-Techn., 42 (1970) 538-540. Eriksson, M., Alderborn, G., The effect of particle fragmentation and deformation on the interparticulate bond formation process during powder compaction. Pharm. Res., 12(7) (1995) 1031-1039. Riepma, K.A., Lerk, C.F., de Boer, A.H., Bolhuis, G.K., Kussendrager, K.D., Consolidation and compaction of powder mixtures: I. Binary mixtures of same particle size fractions of different types of crystalline lactose. Int. J. Pharm., 66 (1990) 47-52. Inglis, C.E., Stresses in a plate due to the presence of cracks and sharp corners. Roy. Inst. Naval Architects Trans., 55 (1913) 219-230. Griffith, A.A., Phenomena of rupture and flow in solids. Phil. Trans. Roy. Soc., A221 (1920) 163-168.

50 51 52

53 54 55 56

57 58

59 60 61

62

63 64

19

Chapter 1 65

Introduction

Mashadi, A.B., Newton, J.M., The characterization of the mechanical properties of microcrystalline cellulose: a fracture meachanics approach. J. Pharm. Pharmacol., 39 (1987) 961-965. Stanley, P., Mechanical strength testing of compacted powders. Int. J. Pharm., 227 (2001) 27-38. Liu, D-M., Influence of porosity and pore size on the compressive strength of porous hydroxyapatite ceramic. Ceramic Int., 23 (1997) 135-139. Johansson, B., Wikberg, M., Ek, R., Alderborn, G., Compression behaviour and compactability of microcrystalline cellulose pellets in relationship to their pore structure and mechanical properties. Int. J. Pharm., 117 (1995) 57-73. Stauffer, D., Introduction to percolation theory. Taylor & Francis, London (1985). Holman, L.E., Leuenberger, H., The effect of varying the composition of binary powder mixtures and compacts on their properties: a percolation phenomenon. Powder Techn., 60 (1990) 249-258. Blattner, D., Kolb, M., Leuenberger, H., Percolation Theory and compactibility of binary powder systems. Pharm. Res., 7(2) (1990) 113-117. Holman, L.E., The compaction behaviour of particulate materials: An elucidation based on percolation theory. Powder Techn., 66 (1991) 265-280. Ehrburger, F., Lahaye, J., Behaviour of colloidal silicas during uniaxial compaction. J. Phys. France, 50 (1989) 1349-1359. Holman, L.E., Leuenberger, H., The significance of slopes of the semilogarithmic relationship between hardness and solid fraction of porous compacts. Powder Techn., 64 (1991) 233-247. Steendam, R., Eissens, A.C., Frijlink, H.W., Lerk, C.F., Plasticisation of amylodextrin by moisture: Consequences for drug release from tablets, Int. J. Pharm., 204(1-2) (2000) 2333. Milln, M., Caraballo, I., Rabasco, A.M., The role of the drug/excipient particle size ratio in the percolation model for tablets. Pharm. Res., 15(2) (1998) 216-220. Larhrib, H., Wells, J.I., Compression speed on polyethylene glycol and dicalcium phosphate tableted mixtures. Int. J. Pharm., 160 (1998) 197-206. Majuru, S., Wurster, D.E., The effect of composition on the tableting indices of binary powder mixtures. Pharm. Dev. & Techn., 2(4) (1997) 313-321. Jetzer, W.E., Compaction characteristics of binary mixtures. Int. J. Pharm., 31 (1986) 201-207. Rubinstein, M.H., Jackson, I.M., Stress relaxation behaviour of compacts of sodium chloride and polyethylene. Int. J. Pharm., 36 (1987) 99-104. Olsson, H., Mattsson, S., Nystrm, C., Evaluation of the effect of addition of polyethylene glycols of differing molecular weights on the mechanical strength of sodium chloride and sodium bicarbonate tablets. Int. J. Pharm., 171 (1998) 31-44.

66 67 68

69 70

71 72 73 74

75

76 77 78 79 80 81

20

Introduction 82

Chapter 1

Mattsson, S., Nystrm, C., Evaluation of strength-enhancing factors of a ductile binder in direct compression of sodium bicarbonate and calcium carbonate powders. Eur. J. Pharm. Sci., 10 (2000) 53-66. Vromans, H., Lerk, C.F., Densification properties and compactibility of mixtures of pharmaceutical excipients with and without magnesium stearate. Int. J. Pharm., 46 (1988) 183-192. Cook, G.D., Summers, M.P., Effect of compression speed on the tensile strength of tablets of binary mixtures containing Aspirin. J. Pharm. Pharmacol., 42 (1990) 462-467. Mattsson, S., Nystrm, C., The use of mercury porosimetry in assessing the effect of different binders on the pore structure and bonding properties of tablets. Eur. J. Pharm. Biopharm., 52 (2001) 237-247. Sheikh-Salem, M., Alkaysi, H., Fell, J.T., The tensile strength of tablets of binary mixtures lubricated with magnesium stearate. Drug Dev. Ind. Pharm., 14(7) (1988) 895903. Larhrib, H., Wells, J.I., Polyethylene glycol and dicalcium phosphate mixtures: effect of tableting pressure. Int. J. Pharm., 159 (1997) 75-83.

83

84 85

86

87

21

Chapter 1

Introduction

22

Tensile strength of tablets containing two materials

Chapter 2

Chapter 2 Tensile strength of tablets containing two materials with a different compaction behaviour

Summary The tensile strength of tablets compressed from binary mixtures is in general not linearly related to the strength of tablets prepared from single materials; in many cases it shows a decreased tensile strength relative to interpolation. This chapter evaluates the compaction behaviour of powder blends consisting of the two model materials used in this study. Sodium chloride and pregelatinised starch are both plastically deforming materials, but have a different densification and relaxation behaviour. The yield pressure of the binary mixtures shows an almost linear relationship. As a result of their lower yield pressure, starch particles yield earlier than sodium chloride particles. The following enclosure prevents some sodium chloride particles from yielding or cracking. The relaxation of the tablets is higher than the relaxation calculated by linear interpolation of the relaxation behaviour of the two pure materials. The difference between the measured porosity expansion and the data obtained by linear interpolation can be considered as a measure for the reduced interparticle bonding. SEM photographs indicate that the reduced interparticle bonding is caused by low adhesive forces. The measured decrease in tensile strength of the tablets is also considered to be the result of reduced interparticle bonding. In this chapter it is shown that a similar relationship exists between the tensile strength reduction and the percentage of starch on the one hand and the extra porosity expansion and the starch percentage on the other hand. Keywords: tablet strength; binary mixture; densification; interparticle bonding; relaxation.

Published in Int. J. Pharm. 203 (2000) 71-79.

23

Chapter 2

Tensile strength of tablets containing two materials

2.1

Introduction

Tablets produced in the pharmaceutical industry commonly consist of more than one component. The mechanical strength of a tablet depends on both formulation and processing parameters. Even with constant processing parameters, the strength of tablets compressed from binary mixtures can often not be predicted from the compaction properties of the starting materials, because interactions between the materials may occur during the compaction process (1). Tablet formation is the result of densification, interparticle bonding and relaxation. Each component has its own compaction behaviour. After rearrangement of the particles and building up of the pressure, the particles of the single materials behave brittle or viscoelastic (2, 3). Additionally, the deformation behaviour depends on other factors like particle size in relation to its critical brittle/ductile transition particle size (4) and compaction velocity (5). After release of the applied load, the tablet shows a certain amount of relaxation before reaching its final porosity. Several investigators have examined the compaction behaviour of binary mixtures. The effects of different parameters, such as compaction behaviour (6), surface area (7) and compaction speed (8, 9), on the final tensile strength of the tablets compressed from blends have been reported. In many cases, the strength of the tablets compressed from the blend is found to be lower than the strength calculated by interpolation of the strength of tablets prepared from the pure materials. The composition of the powder mixture is directly related to a number of tablet properties, such as tablet strength. Furthermore, it is known that the tensile strength of tablets correlates with the porosity according to the Ryshkewitch-Duckworth equation (10). To compare the tensile strength of tablets compressed from different blends of two materials, it is necessary to use volume fractions instead of weight fractions, because it may be expected that the three-dimensional structure of the tablet is more important than the weight ratio of the components (11). Moreover, a constant porosity is necessary to compare the tensile strength of the tablets of the mixtures with different volume percentages. This study was performed to investigate differences between tensile strength of tablets compressed from blends of two materials and the tensile strength of tablets compressed from the individual components. These differences will be related to differences in densification, interparticle bonding and tablet relaxation. The materials chosen (sodium chloride and pregelatinised starch) are two plastically deforming materials, but with different densification and relaxation behaviours (12). The final tablet porosity of sodium chloride is almost completely determined by the

24

Tensile strength of tablets containing two materials

Chapter 2

densification stage, whereas relaxation hardly occurs. In contrast, the ultimate tablet porosity of pregelatinised starch is the effect of both densification and relaxation behaviour (12).

2.2

Materials and methods

The materials used were sodium chloride (glidant-free material, Akzo Nobel, Hengelo, The Netherlands) and pregelatinised potato starch (Paselli WA4, Avebe, Foxhol, The Netherlands). Before use sodium chloride was milled (Moulinex, Birmingham, England) for a short period of time. From both materials the sieve fraction between 106 and 212 m was obtained using an Air Jet Sieve (Alpine, Augsburg, Germany) equipped with USA Standard testing sieves (W.B Tyler, Mentor, OH, USA). Before use the materials were stored at a temperature of 20C and a relative humidity of 60% for a period of at least 4 days. The true density of the 106-212 m particle size distribution was measured by helium pycnometry (Quantachrome, Syosset, N.Y., USA) at a temperature of 20C. The true densities of sodium chloride and pregelatinised potato starch were 2239 kgm-3 and 1486 kgm-3, respectively. The binary mixtures were prepared in volume percentages; mixing was performed with a Turbula mixer, model 2P (W.A. Bachofen, Basle, Switzerland) at 90 rpm for a period of 15 minutes. Tablets (round, flat, diameter 13 mm, weight 500 mg) were compressed on a compaction simulator (ESH, Brierley Hill, UK) with a constant compaction speed of 3 mms-1. The compaction pressures varied between 10 and 350 MPa. The upper punch displacements were sine waves with different amplitudes in order to obtain the different compaction pressures. The lower punch was stationary during compaction. The ejection time was always 10 s. After ejection the tablets were stored for at least 14 hours in a controlled climate chamber (Heraeus, Hanau, Germany) at a temperature of 20C and a relative humidity of 60%. Tablet dimensions were measured with an electronic micrometer (Mitutoyo, Tokyo, Japan) and the tablets were weighed on an analytical balance (Mettler-Toledo, Greifensee, Switzerland). The crushing strength of the tablets was measured with a Schleuniger 4N strength tester (Dr. Schleuniger Productronic, Soloturn, Switzerland). Crushing strengths over 300 N were measured with the compaction simulator (13). The tensile strength (S) of the tablets was calculated according to Fell and Newton (14). The tensile strength was

25

Chapter 2

Tensile strength of tablets containing two materials

calculated for at least 40 tablets with different porosities to determine the Ryshkewitch-Duckworth relation (10) between the tensile strength and the tablet porosity for every binary mixture containing a different percentage pregelatinised starch:

S = S 0 e k

(1)

where S0 is the tensile strength at zero porosity, the porosity and k a constant. k describes the bonding capacity of a single material (15) and indicates the effect of a change in porosity on the tensile strength. From these Ryskewitch-Duckworth relations the tensile strength at a porosity of 15% was calculated for every blend containing a different percentage pregelatinized starch. Yield pressure (Py) of the two materials and the binary mixtures were derived from the force displacement profiles according to Heckel (16, 17). Linear interpolation was performed in the porosity range between 15% and 35%.

2.3 2.3.1

Results and discussion Densification of the binary powder mixtures

During powder densification, the porosity of a powder bed decreases. If the rearrangement of particles of the different blends is considered as a non-dominant factor, the densification can be studied by both the porosity under load and the yield pressure. Fig. 2.1a reflects the porosity under load of mixtures of sodium chloride with the different percentages pregelatinised starch. The porosity under load, using compaction pressures to obtain tablets with a final porosity of 15% after relaxation, decreases. Moreover, the slope decreases upon increasing pregelatinised starch concentration. Obviously, the largest influence of the pregelatinised starch particles on the densification of the sodium chloride particles occurs at low starch fractions. Fig. 2.1b shows the yield pressure of sodium chloride, pregelatinised starch and their binary mixtures. The yield pressure of the two single materials illustrates the difference in densification behaviour. As reported previously, sodium chloride has a significantly higher yield pressure than pregelatinised starch (12, 18). The high yield pressure of sodium chloride illustrates its difficult densification. The figure shows that the second component, pregelatinised starch, with the lowest yield pressure enhances the densification of the powder blends. This results in a lower porosity under pressure for the binary mixtures than for pure sodium chloride

26

Tensile strength of tablets containing two materials

Chapter 2

0.16 Porosity under load (-) 0.12 0.08 0.04 0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 2.1a. Porosity under load of different binary mixtures of sodium chloride and pregelatinised starch for obtaining tablets with a final porosity of 15%.
120 Yield pressure (MPa)

80

40

0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 2.1b. Yield pressure of sodium chloride, pregelatinised starch and their mixtures (symbols) and the yield pressure of the binary mixtures calculated by linear interpolation (dotted line).

If the two components densify independently, the yield pressure of a mixture (Py,mix) can be described by linear interpolation (dotted line, Fig. 2.1b) of the yield pressures of each of the two materials (PNaCl, Pstarch) :

Py ,mix = PNaCl NaCl + Pstarch starch

(2)

with NaCl and starch being the volume fractions of sodium chloride and pregelatinised starch in the blends. The explanation of the small deviations from linearity of the data

27

Chapter 2

Tensile strength of tablets containing two materials

in Fig. 2.1b with eq. 2 can be found from SEM-photographs of the cross-sections of the tablets (Fig. 2.2a-2.2c).
Fig. 2.2a

S
Fig. 2.2b

S
Fig. 2.2c

Fig. 2.2. Cross-section of a tablet compressed from a binary mixture of sodium chloride and (a) 10% pregelatinised starch (b) 20% pregelatinised starch and (c) 40% pregelatinised starch (S = starch particles, N = sodium chloride particles).

28

Tensile strength of tablets containing two materials

Chapter 2

All visual pregelatinised starch and sodium chloride particles in binary mixtures containing 10% (Fig. 2.2a) or 20% pregelatinised starch (Fig. 2.2b) have yielded and/or are cracked. The SEM picture of tablets compressed from a blend containing 40% pregelatinised starch (Fig. 2.2c) shows that all the starch particles are deformed. However, a number of the sodium chloride particles still have their original shape and are not cracked, even though the compaction pressure was higher than the yield pressure. Obviously, the starch particles, which are arranged around the sodium chloride particles, prevent that these sodium chloride particles are affected to a large extent by the compaction process. So, the sodium chloride particles do not yield or crack. This explains that the measured yield pressures of the blends are somewhat lower than those described by eq. 2. The difference in yield pressure between the two components is responsible for this phenomenon. In a binary mixture of sodium chloride and pregelatinised starch under the given conditions, a sodium chloride particle will only yield if it is in direct contact with another sodium chloride particle. However, the majority of the particles in the blends do yielded during the densification of the powder. Therefore, the deviations from linearity are small. Although the particle size of sodium chloride is larger than its critical particle diameter of 35 m (4), the particles possess a low degree of fragmentation at high pressures (19). Under high pressures and close packing of the particles, there is hardly any possibility for the particles to rearrange. It may therefore be assumed that this small fragmentation only has a minor influence on the eventual tensile strength. 2.3.2 Relaxation of the tablets

The relaxation of tablets can be quantified by an increase in tablet height or volume after compression (20). Van der Voort Maarschalk et al. (13) have suggested that tablet relaxation is the result of two phenomena: a decrease in material density and an increase in tablet porosity. In general, the compression-induced increase in material density is fully reversible. The increase in tablet porosity occurs during and after release of the pressure from the tablet. The porosity expansion () is given by:

= min

(3)

where is the tablet porosity after relaxation and min is the porosity under load. min is zero if the porosity under load was negative as an effect of the increase of material density. Just as previously found for single materials (15), the porosity expansion for all the different blends of sodium chloride and pregelatinised starch was found to be independent of the compaction pressures used.

29

Chapter 2

Tensile strength of tablets containing two materials

Fig. 2.3 depicts the relationship between the percentage pregelatinised starch and the porosity expansion for tablets compressed from blends of sodium chloride and pregelatinised starch. In contrast to tablets compressed from pure pregelatinised starch, tablets compressed from sodium chloride show only a small porosity expansion. Fig. 2.3 also shows that the porosity expansion of tablets compressed from the blends is not linearly related with the volume percentage pregelatinised starch, but shows a sharp increase at low fractions of pregelatinised starch. The effect of the addition of pregelatinised starch particles is higher than expected from linear interpolation of the porosity expansion.
0.16 Porosity expansion (-) 0.12 0.08 0.04 0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 2.3. Porosity expansion of tablets compressed from sodium chloride, pregelatinised starch and their mixtures (symbols) and the porosity expansion calculated by linear interpolation (dotted line).

Under load, the particles of sodium chloride and pregelatinised starch are closely packed. During tablet relaxation, the tablet volume increases, resulting in an increase in size and number of pores. The formation of extra pores consequently leads to the formation of more free particle surfaces, which is an effect of partial or complete disconnection of particles. In a tablet compressed from a blend of sodium chloride and pregelatinised starch particles, there are three different types of particle-particle bonding that can, partially or completely, be disconnected: 1) the cohesive bonding between sodium chloride particles 2) the cohesive bonding between pregelatinised starch particles 3) the adhesive bonding between sodium chloride and pregelatinised starch particles.

30

Tensile strength of tablets containing two materials

Chapter 2

The SEM-photographs (Fig. 2.2a-2.2c), especially the one of tablets compressed from the blend of sodium chloride and 40% pregelatinised starch (Fig. 2.2c), show that pregelatinised starch particles situated around the individual sodium chloride particles have an imprint of the enclosed sodium chloride particles. This implies that pregelatinised starch particles, which were in contact with sodium chloride particles under load, have been detached. This observation is a strong indication for the fact that the adhesion forces between the sodium chloride and pregelatinised starch particles are smaller than the two different cohesion forces. The observation of smaller adhesion forces can be explained by the fact that sodium chloride and pregelatinised starch have a different type of interparticle bonding (21). 2.3.3 Tablet strength

Fig. 2.4 shows the tensile strength of tablets with a porosity of 15% compressed from different binary mixtures of sodium chloride and pregelatinised starch as a function of the volume percentage pregelatinised starch. The figure shows that the tensile strength of tablets compressed from 100% pregelatinised starch is higher than that of tablets containing 100% sodium chloride. Furthermore, the strengths are lower than predicted by interpolation.
5 Tensile strength (MPa) 4 3 2 1 0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 2.4. Tensile strength of tablets compressed from sodium chloride, pregelatinised starch and their mixtures (symbols) and the tensile strength calculated by linear interpolation (dotted line).

Strength reduction can be considered as the result of reduction in interparticle bonding. In line with our conclusions from the SEM-photographs, the decreased

31

Chapter 2

Tensile strength of tablets containing two materials

tensile strength of tablets compressed from the blends indicates that the adhesive forces between the sodium chloride particles and the pregelatinised starch particles are smaller than the cohesive forces between particles of the pure materials. The reduction of interparticle bonding inside tablets compressed from different volume percentages sodium chloride and pregelatinised starch is reflected in Fig. 2.5 as the difference between the measured tensile strength and the calculated tensile strength (dotted line, Fig. 2.4).
- Difference tensile strength (MPa) 3 2.4 1.8 1.2 0.6 0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 2.5. Difference between the measured tensile strength and calculated tensile strength of tablets compressed from sodium chloride with different volume.

0.06 Extra porosity expansion (-)

0.04

0.02

0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 2.6. Extra porosity expansion of tablets compressed from sodium chloride with different volume percentages pregelatinised starch.

32

Tensile strength of tablets containing two materials

Chapter 2

A decreased tensile strength, caused by a reduction of interparticle bonding, is related to a larger relaxation of the tablets (15, 22, 23). Therefore, the extra porosity expansion between the measured porosity expansion and the porosity calculated by linear interpolation (Fig. 2.3) is suggested to be a measure for the reduced interparticle bonding. This extra porosity expansion is depicted in Fig. 2.6. Both the decrease in tensile strength of tablets compressed from sodium chloride and pregelatinised starch (Fig. 2.5) and the extra porosity expansion (Fig. 2.6) are a reflection of the reduced interparticle bonding inside the tablets. Comparison of the figures shows that an almost similar relationship was found between the decrease in tensile strength and the percentage pregelatinised starch on the one hand and the extra porosity expansion of the tablets and the percentage pregelatinised starch on the other hand. Figures 2.5 and 2.6 have been combined into Fig. 2.7. This figure indeed suggests such an interrelation of the extra porosity expansion, the reduced interparticle bonding and the decrease in tensile strength.
- Difference tensile strength (MPa) 3

0 0 0.02 0.04 0.06 Extra porosity expansion (-)

Fig. 2.7. Difference in tensile strength as a function of the extra porosity expansion.

In conclusion, the yield pressures of blends of sodium chloride with different percentages pregelatinised starch can be calculated by linear interpolation of the yield pressures of the two single materials. The porosity expansion of tablets compressed from the blends is higher than calculated from linear interpolation of the porosity expansion of tablets compressed from the pure materials. Both the extra porosity expansion and the decrease in tensile strength of the tablets compressed from the binary blends are related and caused by a reduction in interparticle bonding. Using the interparticle bonding reduction, caused by adhesion forces that are lower than the

33

Chapter 2

Tensile strength of tablets containing two materials

cohesion forces, it is possible to predict the relaxation behaviour of the mixture using the properties of the individual components.

2.4 k Py Py,mix PNaCl Pstarch S S0 min NaCl starch

Nomenclature constant, bonding capacity yield pressure yield pressure of binary mixture yield pressure of sodium chloride yield pressure of pregelatinised starch tensile strength tensile strength at zero porosity porosity porosity under load porosity expansion volume fraction of sodium chloride volume fraction of pregelatinised starch

2.5
1

References
Vromans, H., Lerk, C.F., Densification properties and compactibility of mixtures of pharmaceutical excipients with and without magnesium stearate. Int J. Pharm., 46 (1988) 183-192. Duberg, M., Nystrm, C., Studies on direct compression of tablets: XVII. Porositypressure curves for the characterization of volume reduction mechanisms in powder compresssion. Powder Techn., 46 (1986) 67-75. Van der Voort Maarschalk, K., Bolhuis, G.K., Improving properties of materials for direct compression, Part 1. Pharm. Techn. Europe, 10(9) (1998) 30-35. Roberts, R.J., Rowe, R.C., Kendall, K., Brittle-ductile transitions in die compaction of sodium chloride. Chem. Eng. Sci., 44(8) (1989) 1647-1651. Roberts, R.J., Rowe, R.C., The effect of punch velocity on the compaction of a variety of materials. J. Pharm. Pharmacol., 37 (1985) 377-384. Sheikh-Salem, M., Fell, J.T., Compaction characteristics of mixtures of materials with dissimilar compaction mechanisms. Int. J. Pharm. Tech. & Prod. Mfr., 2(1) (1981) 19-22. Adolfsson, ., Caramella, C., Nystrm, C., The effect of milling and addition of dry binder on the interparticulate bonding mechanisms in sodium chloride tablets. Int. J. Pharm., 160 (1998) 187-195. Cook, G.D., Summers, M.P., Effect of compression speed on the tensile strength of tablets of binary mixtures containing aspirin. J. Pharm. Pharmacol., 42 (1990) 462-467.

3 4 5 6 7

34

Tensile strength of tablets containing two materials 9

Chapter 2

Garr, J.S.M., Rubinstein, M.H., The effect of rate of force application on the properties of microcrystalline cellulose and dibasic calcium phosphate mixtures. Int. J. Pharm., 73 (1991) 75-80. Duckworth, W.H., Discussion of Ryshkewitch paper by Winston Duckworth. J. Am. Ceram. Soc., 36 (1953) 68. Bangudu, A.B., Pilpel, N., Tensile strength of paracetamol and Avicel powders and their mixtures. J. Pharm. Pharmacol., 36 (1984) 717-722. Van der Voort Maarschalk. K., Zuurman, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., Stress relaxation of compacts produced from viscoelastic materials. Int. J. Pharm., 151 (1997) 27-34. Van der Voort Maarschalk, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., The effect of viscoelasticity and tabletting speed on consolidation and relaxation of a viscoelastic material. Eur. J. Pharm. Biopharm., 42 (1996) 49-55. Fell, J.T., Newton, J.M., The tensile strength of lactose tablets. J. Pharm. Pharmacol., 20 (1968) 657-658. Van der Voort Maarschalk, K., Zuurman, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., Porosity expansion of tablets as a result of bonding and deformation of particulate solids. Int. J. Pharm., 140 (1996) 185-193. Heckel, R.W., Density-pressure relationships in powder compaction. Trans. Met. Soc. AIME., 221 (1961) 671-675. Heckel, R.W., An analysis of powder compaction phenomena. Trans. Met. Soc. AIME., 221 (1961) 1001-1008. Duberg, M., Nystrm, C., Studies on direct compression of tablets: XII. The consolidation and bonding properties of some pharmaceutical compounds and their mixtures with Avicel 105. Int. J. Pharm. Tech. & Prod. Mfr., 6(2) (1985) 17-25. Alderborn, G., Pasanen, K., Nystrm, C., Studies on direct compression of tablets. XI. Characterization of particle fragmentation during compaction by permeametry measurements of tablets. Int. J. Pharm., 23 (1985) 79-86. Doelker, E., Comparative compaction properties of various microcrystalline cellulose types and generic products. Drug Dev. Ind. Pharm., 19 (1993) 2399-2471. Adolfsson, ., Gustafsson, C., Nystrm, C., Use of tablet tensile strength adjusted for surface area and mean interparticulate distance to evaluate dominating bonding mechanisms. Drug Dev. Ind. Pharm., 25(6) (1999) 753-764. Rees, J.E., Tsardaka, K.D., Some effects of moisture on the viscoelastic behaviour of modified starch during powder compaction. Eur. J. Pharm. Biopharm., 40 (1994) 193197. Zuurman, K., Van der Voort Maarschalk, K., Bolhuis, G.K., Effect of magnesium stearate on bonding and porosity expansion of tablets produced from materials with different consolidation properties. Int. J. Pharm., 179 (1999) 107-115.

10 11 12

13

14 15

16 17 18

19

20 21

22

23

35

Chapter 2

Tensile strength of tablets containing two materials

36

Pore formation in tablets containing two materials

Chapter 3

Chapter 3 Pore formation in tablets compressed from binary mixtures as a result of deformation and relaxation of particles

Summary This chapter describes the internal structure of tablets compressed from binary mixtures of sodium chloride and pregelatinised starch. The minimum particle diameter of pregelatinised starch inside tablets compressed from mixtures was calculated from the difference between the initial pore size distribution and the pore size distribution after removal of the starch particles by burning. Subsequently, the tablets were carefully crushed. These powders, consisting of almost only sodium chloride particles, were measured by laser diffraction. It was found that the diameter of the sodium chloride particles hardly changed, whereas the minimum diameter of starch particles strongly decreased during the compaction process. As an effect of the difference in yield pressure, the harder sodium chloride particles cause deformation of the softer starch particles, resulting in a change in particle shape. The pore size distribution of tablets compressed from mixtures of sodium chloride and starch is typically that of viscoelastic materials; the larger pores (> 5 m) change, while the small pores stay constant in number and size. The median pore diameter in tablets compressed from the mixtures is higher than the median pore diameter in tablets compressed from the pure materials. This chapter shows that the formation of large pores is the result of the extra porosity expansion of tablets compressed from binary mixtures of sodium chloride and pregelatinised starch. Keywords: binary mixture; particle size distribution; pore size distribution; relaxation; tablet.

Published in Eur. J. Pharm. Sci. 15 (2002) 171-177.

37

Chapter 3

Pore formation in tablets containing two materials

3.1

Introduction

The properties of a tablet are determined by those of the starting materials and by the processing conditions (1, 2). However, they are more easily understood in terms of the resulting structure of the tablet. This can be described by the volume fractions of the components, of their particle sizes, of porosity and other experimental parameters An often applied parameter to express the structure of a tablet is its porosity. Commonly, the tablet porosity is calculated from the weight and volume of the tablet and the true densities of the ground materials. However, to characterise the internal structure of a tablet it is better to describe porosity not only in terms of the total volume fraction of the pores, but also in terms of the pore size distribution (3, 4). Papers presented so far determine the pore size distribution of tablets containing one excipient, or do not make a distinction between the different excipients used. It is, however, reasonable to believe that the differences in properties and the (particle size) distributions of the excipients are as important as the pore size distribution. The goal of this study is to determine the size distribution of pores and of excipients in tablets compressed from a mixture of two materials (sodium chloride and pregelatinised starch). Furthermore, the present paper attempts to give more information regarding the origin of pore formation in tablets made from a powder blend.

3.2 3.2.1

Materials and methods Materials

The materials used were sodium chloride (glidant-free material, Akzo Nobel, Hengelo, The Netherlands) and pregelatinised potato starch (Paselli WA4, Avebe, Foxhol, The Netherlands). The sodium chloride powder was milled for a short time in a coffee mill (Moulinex, Birmingham, UK) to obtain smaller particles. The 106-212 m sieve fractions of both materials were used in the experiments. Sieving was performed by using USA Standard sieves (W.B. Tyler Inc., Mentor, OH, USA) on an Alpine Air Jet Sieve (Alpine, Augsburg, Germany). Before the binary mixtures were prepared, the powders were separately conditioned at 20C and 60% relative humidity (R.H.) for at least 4 days. True particle density of the single materials was obtained using helium pycnometry measurements (Quantachrome, Syosset, NY, USA). The true densities of sodium chloride and pregelatinised starch were 2239 and 1486 kgm-3, respectively. The apparent particle density of the powder mixtures was

38

Pore formation in tablets containing two materials

Chapter 3

calculated from the true densities of the single materials. The powder mixtures prepared had pregelatinised starch concentrations between 2.5 and 80 % on a volume basis. Tablets compressed from these different mixtures were used to determine the changes in pore size distribution as accurately as possible. Mixing was performed with a Turbula mixer (W.A. Bachofen, Basle, Switzerland) at 90 rpm for a period of 15 min. 3.2.2 Tablets

Tablets produced in this study were cylindrical, had a diameter of 13 mm, and a mass of approximately 500 mg. Compaction was performed with a compaction simulator (ESH, Brierley Hill, UK) using an average punch speed of 3 mms-1 at 20C and 60% relative humidity. The die and punches were prelubricated with magnesium stearate. The upper punch displacements were sine waves with different amplitudes in order to obtain the densification necessary to gain tablets with a final porosity of 15%. The lower punch was stationary during compaction and the ejection time was always 10 s. The tablets were stored for at least 14 hours in a controlled climate chamber (Heraeus, Hanau, Germany) at 20C and 60% RH. To calculate the final porosity, tablet dimensions were measured with an electronic micrometer (Mitutoyo, Tokyo, Japan) and tablets weights were determined with an analytical balance (Mettler-Toledo, Greifensee, Switzerland). 3.2.3 Pore size distribution

The pore size distributions of the tablets with a porosity of 15% ( 0.5%) were determined by mercury intrusion porosimetry (Micromeritics, Model Autopore 9220, Norcross, GA, USA). Three tablets were used for the measurement of the pore size distribution of each mixture. Low pressure measurements between 7 and 171 kPa were performed to measure pores with a diameter between 9 and 220 m. The pressures used for the high pressure measurements varied from 0.2 to 14 MPa, which correspond to pore diameters in the range of 7 0.1 m. It was assumed that the surface tension of mercury was 480 mNm-1 and the contact angle of mercury with the materials was assumed to be 140 in all cases.

39

Chapter 3

Pore formation in tablets containing two materials

3.2.4

Heat treatment

To determine the particle size distribution of pregelatinised starch inside the tablet structure, three tablets of each composition with a porosity of 15% ( 0.5%) were placed in a muffle furnace (Naber Industrieofenbau, Bremen, Germany). Before the tablets were placed in the oven, they were dried in a dessicator (containing silica) for 24 hours to remove the adsorbed water. From literature it is known that the melting temperature of sodium chloride is 800C, while the self-ignition temperature of pregelatinised starch is 300C (5). Besides, pregelatinised starch starts to degrade at much lower temperatures. The temperature trajectory used was 250C for the first six hours, after which the temperature was increased to 280C for another 60 hours. After the burning period, the tablets were cooled down in a dessicator and conditioned at 20C and 60% R.H. The dimensions and weights of the tablets after burning were measured with the same method as the one used for untreated tablets. The pore size distributions of these tablets were measured by mercury intrusion porosimetry under the same conditions as described in sec. 3.2.3. 3.2.5 Laser diffraction

To measure the occurrence of a change in particle size of sodium chloride during compaction, tablets were carefully crushed with a spatula after the heat treatment. To see whether all primary particles were separated, the powder was visually checked under a microscope (Olympus Optical Co., Olympus BX 50, Tokyo, Japan). The particle size distribution of these powders was measured by laser diffraction (Sympatec, Helos KA/LA, Clausthal-Zellerfeld, Germany).

40

Pore formation in tablets containing two materials

Chapter 3

3.3 3.3.1

Results and discussion Pregelatinised starch particle size distribution

A method to investigate the solid structure formed by pregelatinised starch particles during compaction is the removal of these particles. In tablets compressed from sodium chloride and pregelatinised starch, the low degradation temperature of pregelatinised starch (compared with sodium chloride) enables the removal of the starch particles by burning them out. It leaves the sodium chloride particles behind inside the tablet structure. Table 3.1 lists the percentage pregelatinised starch that was removed and shows that it was possible to remove at least 92% of the initially present amount of pregelatinised starch by the heat treatment. The pores created by elimination of the pregelatinised starch particles will be filled up by air. As a consequence, the pore size distribution in the tablets will change. The pore size distribution after removal will not only consist of the new pores that were formed by removal of the pregelatinised starch particles, but also of the initial pores when all particles were present. The initial and new pore size distribution can both be measured by mercury intrusion porosimetry.
Table 3.1. Percentage of total initial amount of pregelatinised starch present inside the tablet that was removed from the tablet by the heat treatment. v/v % NaCl / pregelatinised starch 97.5/2.5 95/5 92.5/7.5 90/10 85/15 80/20 % removed pregelatinised starch 99 98 99 95 92 too weak for further measurement

To measure the pore size distribution after removal of the pregelatinised starch particles, it is required that the initial shape of the tablet is retained. The sodium chloride particles need to have sufficient coherence to preserve the dimensions of the tablet. After the heat treatment, tablets with an initial sodium chloride concentration in the blend of at least 85% (Table 3.1) were still strong enough to allow measurement of the pore size distribution by mercury intrusion porosimetry.

41

Chapter 3

Pore formation in tablets containing two materials

Furthermore, the heat treatment must not harm the structure formed by sodium chloride particles. Fig. 3.1 shows the pore size distributions of 100% sodium chloride tablets with a porosity of 15% before and after the heat treatment.
0.006 Pore size distribution (ml m-1 g-1)

0.004

0.002

0 100

10 Pore diameter (m)

Fig. 3.1. Pore volume size distribution of 100% sodium chloride tablets with a porosity of 15% () before and () after heat treatment.

There are hardly any differences between both distributions for pore diameters larger than 8 m. For pores smaller than 8 m, a maximum deviation of 0.001 mlm-1g-1 was found between the pore size distributions. For this reason, for diameters smaller than 8 m, a judgement about the particle size distribution of starch in tablets compressed from the binary mixtures is only possible if the effect is higher than 0.001 mlm-1g-1. Sodium chloride appears to behave inertly with respect to the moderate temperature increase. Therefore, it is reasonable to assume that any alteration in pore size distribution after burning out the starch is due to the starch removal. This alteration is a representation of the distribution of the starch particles in the tablet. Fig. 3.2a displays the pore size distribution of tablets containing different ratios of sodium chloride to pregelatinised starch. These graphs represent the volume distribution of the pores in tablets before the heat treatment and will be discussed further in sec. 3.3.3. Before the heat treatment, all tablets had a porosity of 15%. The measured pore size distribution of tablets after burning out the pregelatinised starch particles is depicted in Fig. 3.2b. As an effect of the heat treatment, a higher pore volume has been created inside the tablets with increasing amounts of pregelatinised starch. A comparison of Fig. 3.2a with Fig. 3.2b shows that the higher pore volume is mainly created by pores with a diameter between 10 and 100 m.

42

Pore formation in tablets containing two materials

Chapter 3

0.006 Pore size distribution (ml m-1 g-1)

0.004

0.002

0 100

10 Pore diameter (m)

Fig. 3.2a. Pore volume size distributions of tablets compressed from () sodium chloride, () a binary mixture of sodium chloride and 7.5% pregelatinised starch and () a binary mixture of sodium chloride and 15% pregelatinised starch.

0.008 Pore size distribution (ml m-1 g-1) 0.006 0.004 0.002 0 100

10 Pore diameter (m)

Fig. 3.2b. Pore volume size distributions of tablets after burning out of the pregelatinised starch particles. Tablets compressed from blends of sodium chloride and pregelatinised starch concentrations of () 2.5%, () 5%,() 7.5%, (*) 10% and (+) 15%.

The intrusion of mercury into these new holes can be visualised by calculating the difference between the pore size distribution before and after the heat treatment (Fig. 3.3). This figure shows that the pores that were originally occupied by starch have a typical size between 8 and 20 m, while a smaller fraction pores is in the size range between 20 and 100 m. These data reveal that the particle size of starch has been

43

Chapter 3

Pore formation in tablets containing two materials

reduced from between 106 and 212 m to the values indicated. Pregelatinised starch is known deform plastically (6). The volumes of the individual starch particles do not alter upon compression. This means that fracture of the starch particles is not the explanation for this apparent reduction of the starch particle size. In analogy with what has been presented in Ch. 2, the yield strength difference between sodium chloride and starch implies that the starch undergoes severe deformation upon compression. The result of this deformation is that the starch particles are flattened into pan cakes. It is the thickness of this pan cake that determines the pressure for mercury intrusion, because this is the radius of curvature that determines the pressure for intrusion. Apparently, the thickness of the flat particles is in the order of 8 to 20 m and can be seen as the minimum particle diameter of starch.
0.004 Particle size distribution (ml m-1 g-1) 0.003 0.002 0.001 0 -0.001 100

10 Particle diameter (m)

Fig. 3.3. Particle volume size distribution of pregelatinised starch particles inside tablets compressed from blends of sodium chloride and different concentrations pregelatinised starch as calculated from pore volume size distributions before and after the heat treatment. Symbols as in Fig. 3.2.

3.3.2

Sodium chloride particle size distribution

After the heat treatment, the tablets listed in Table 3.1, which for the most part contain sodium chloride, were carefully crushed under low forces with a spatula, preventing fragmentation of primary particles. To ensure that most of the particles were well separated from each other, the powders were examined under a microscope. The particle size distribution of these powders was determined by laser diffraction measurements. The results of the different measurements are displayed in Fig. 3.4. The particle size distribution of sodium chloride before compression is represented by

44

Pore formation in tablets containing two materials

Chapter 3

the dotted line and has a mean particle size of 170 m. The size distribution of the heat-treated and subsequently crushed tablets represent the particle size distribution of sodium chloride, since most of the pregelatinised starch was burned out (Table 3.1). These particle size distributions are almost equal to the initial particle size distribution of the pure sodium chloride before compaction (Fig. 3.4). The measured particles with a diameter around 10 m are suggested to be residual starch particles and some sodium chloride particle fragments (because sodium chloride shows very limited cracking and fragmentation too).
0.4 Particle size distribution (ml m-1 g-1) 0.3 0.2 0.1 0 1000

100

10

Particle diameter (m)

Fig. 3.4. Particle volume size distribution of sodium chloride before compression (dotted line) and particles of burned and carefully crushed tablets compressed from () sodium chloride and 2.5% pregelatinised starch and () sodium chloride and 10% pregelatinised starch, respectively.

A comparison of the curves depicted in Fig. 3.4 proves that the differences in particle size before and after compression is small. Realising that sodium chloride is a plastically deforming material (7) and following the discussion in sec. 3.3.1, it can be concluded that the shape of the sodium chloride particles remains essentially the same when compressed in a mixture with pregelatinised starch. The reason of the large deformation of starch and the limited deformation of sodium chloride can be attributed to the differences in rigidities described in Ch. 2. Pregelatinised starch is known to be a soft material with a low yield pressure, whereas sodium chloride is a tough material with a high yield pressure. It must be kept in mind that the three different particle sizing techniques used in this study (sieving, mercury intrusion porosimetry and laser diffraction) are based on different principles and produce different types of particle size distributions. The

45

Chapter 3

Pore formation in tablets containing two materials

different shape factors of the sodium chloride and pregelatinised starch particles also have different influences on the results of the measurements. However, the large difference found between the particle size of sodium chloride and pregelatinised starch gives a clear indication of the different particle sizes inside the tablet structure and the origin of this difference. 3.3.3 Pore size distribution

As displayed in Fig. 3.2a, the pore size distributions of tablets with a porosity of 15% compressed from different mixtures of sodium chloride and pregelatinised starch were measured by mercury intrusion porosimetry. For the sake of clarity, not all graphs were implemented in Fig. 3.2a. The pore size distributions shown are given to demonstrate the typical shapes of the measured pore size distributions. Because the excipients used are not microporous and no pores in the sub-micron range from 1 to 300 nm were present (8), the measured pores can be regarded as interparticle pores. To elucidate the effects that occur when the composition of the mixture changes, the pore size distributions were arbitrarily divided into three parts: small pores (< 5 m), medium pores (5-10 m) and large pores (> 10 m).
Percentage pore size distribution -1 -1 (% ml m g ) 100 60 90 50 80 40 70 30 20 10 0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 3.5a. Percentages pore volume of pores with diameters () < 5m, () 5-10 m and () > 10 m as function of the tablet composition.

Fig. 3.5a gives the three pore types as a function of the tablet composition. The pore fraction smaller than 5 m hardly depends on composition. Larger differences are found for pores with a diameter above 5 m. The medium-sized pores have a minimum for tablets containing 20% pregelatinised starch. For the large pores a

46

Pore formation in tablets containing two materials

Chapter 3

reverse effect can be seen: the percentage of the total pore volume increases to a maximum for tablets compressed from 20% pregelatinised starch and decreases at higher concentrations. That the larger pores change in size, while the smaller pores show only a limited change in size and number is well documented for the compaction behaviour of tablets consisting of a single viscoelastic material (9, 10, 11). The decrease of the larger pores is coupled to an increase in compaction pressure. Because different compaction pressures are needed to produce tablets of the different mixtures with a porosity of 15% (results not shown), the change in pore size distribution of tablets compressed from a blend of two viscoelastic materials cannot simply be compared with that of a tablet compressed from a single viscoelastic material. However, a limited change in size of the smaller pores is found in both cases. Particles of viscoelastic materials, like sodium chloride and pregelatinised starch, show no or a limited amount of fragmentation causing no extra small pores. To compare the change in pore size distribution between tablets of the different blends, the median pore diameter was calculated for all binary mixtures. Because of their larger volume, the medium and large pores influence the median pore diameter much more than the small pores. The median pore diameter of tablets compressed from binary blends of sodium chloride and pregelatinised starch is given in Fig. 3.5b. The change of the median pore diameter is clearly in accordance with the described transitions of the pore size distributions in Fig. 3.5a.
Median pore diameter ( m) 13 11 9 7 5 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 3.5b. Median pore diameter of tablets compressed from sodium chloride, pregelatinised starch and their mixtures.

47

Chapter 3

Pore formation in tablets containing two materials

3.3.4

The formation of large pores

Every compaction cycle of the production of a tablet consists of a densification and a relaxation step. The combination of both steps results in the final tablet porosity. The porosity under pressure is a measure for the densification of the powder. The relaxation can be reflected by the porosity expansion, which is the difference between the final tablet porosity and the porosity under pressure. The porosity expansion of tablets compressed from single materials is considered to be independent on the compaction pressure (12). This pressure independence was also found to be valid for all the different mixtures of sodium chloride and pregelatinised starch (results not shown). Fig. 3.6a shows the porosity expansion as a function of the percentage pregelatinised starch for tablets compressed from blends of sodium chloride and pregelatinised starch.
0.16 Porosity expansion (-) 0.12 0.08 0.04 0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 3.6a. Porosity expansion of tablets compressed from sodium chloride with different concentrations pregelatinised starch () and the porosity expansion of the binary mixtures calculated by linear interpolation (dotted line).

In Fig. 3.6a, it is shown that 100% sodium chloride tablets have a small porosity expansion of 1.5%. This implies a relatively low densification of the powder bed to 13.5% porosity under pressure to gain tablets with a final porosity of 15%. Fig. 3.6a also shows that all tablets compressed from mixtures have a higher relaxation than expected from linear interpolation using single component data. In order to be able to reach a final porosity of 15%, these tablets need a larger densification than is necessary for pure sodium chloride tablets. According to Bockstiegels theory (9, 10), this increased densification of a powder bed containing plastic materials results in

48

Pore formation in tablets containing two materials

Chapter 3

smaller pores between particles. However, the median pore diameter reflected in Fig. 3.5b of tablets compressed from the mixtures is higher than the median pore diameter of either 100% sodium chloride or 100% pregelatinised starch tablets. Therefore, it is suggested that relaxation rather than densification is the cause of the creation of larger pores inside tablets compressed from mixtures of pregelatinised starch and sodium chloride.
Extra porosity expansion (-) 0.04 0.03 0.02 0.01 0 0 20 40 60 80 100 v/v % Pregelatinised starch

Fig. 3.6b. Extra porosity expansion of tablets compressed from sodium chloride, pregelatinised starch and their mixtures.

To support this suggestion, the extra porosity expansion as defined previously in Ch. 2 has been calculated as a function of the tablets compositions (Fig. 3.6b). This extra porosity formation is a result of the compressed powder being a mixture rather than a single compound. The observation that there is extra porosity expansion implies that extra pores are formed or that existing pores increase in size as an effect of porosity expansion (i.e. tablet relaxation). The maximum in extra porosity expansion (Fig. 3.6b) and the median pore diameter (Fig. 3.5b) is found for the mixture containing 20% starch. This corresponds to a critical composition in which apparently the resistance against extra pore formation is the lowest. The effect of tablet structure and pore structure on the tablet strength will be the subject of the following chapter.

49

Chapter 3

Pore formation in tablets containing two materials

Median pore diameter ( m)

13 11 9 7 5 0 0.01 0.02 0.03 0.04 Extra porosity expansion (-)

Fig. 3.7. Median pore diameter as a function of the extra porosity expansion.

Fig. 3.7 shows that there is a direct relation between the extra porosity expansion and the median size of pores in the tablets. It is clear that extra relaxation results in formation of larger pores. Therefore, it is concluded that relaxation determines the final pore size distribution in tablets compressed from powder blends much more than compression.

50

Pore formation in tablets containing two materials

Chapter 3

3.4
1

References
Rowe, R.C., Roberts, R.J., Mechanical properties. In: Alderborn, G. and Nystrm, C. (Eds.), Pharmaceutical Powder Compaction Technology. Marcel Dekker, New York (1996) 283-322. Van der Voort Maarschalk, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., The effect of viscoelasticity and tabletting speed on consolidation and relaxation of a viscoelastic material. Eur. J. Pharm. Biopharm., 42 (1996) 49-55. Juppo, A.M., Porosity parameters of lactose, glucose and mannitol obtained by mercury porosimetry. Int. J. Pharm., 129 (1996) 1-12. Riippi, M., Yliruusi, J., Niskanen, T., Kiesvaara, J., Dependence between dissolution rate and porosity of compressed erythromycin acistrate tablets. Eur. J. Pharm. Biopharm., 46 (1998) 169-175. Weast, R.C., Astle, M.J., CRC Handbook of chemistry and physics, 60th edition. CRC Press Inc., Boca Raton, Florida (1979-1980). Rees, J.E., Tsardaka, K.D., Some effects of moisture on the viscoelastic behaviour of modified starch during powder compaction. Eur. J. Pharm. Biopharm., 40 (1994) 193197. Duberg, M., Nystrm, C., Studies on direct compression of tablets: XII. The consolidation and bonding properties of some pharmaceutical compounds and their mixtures with Avicel 105. Int. J. Pharm. Tech. & Prod. Mfr., 6(2) (1985) 17-25. Westermarck, S., Juppo, A.M., Kervinen, L., Yliruusi, J., Pore structure and surface area of mannitol powder, granules and tablets determined with mercury porosimetry and nitrogen adsorption. Eur. J. Pharm. Biopharm., 46 (1998) 61-68. Bockstiegel, G., Relations between pore structure and densification mechanism in the compacting of iron powders. I. Compacting properties in relation to the pore structure inside and between powder particles. Int. J. Powder Metall., 2(4) (1966) 13-26. Bockstiegel, G., Relations between pore structure and densification mechanism in the compacting of iron powders. II. Theoretical considerations about the change of pore size distribution in compacting. Int. J. Powder Metall., 3(1) (1967) 29-37. Vromans, H., de Boer, A.H., Bolhuis, G.K., Lerk, C.F., Kussendrager, K.D., Bosch, H., Studies on tableting properties of lactose. Part. 2. Consolidation and compaction of different types of crystalline lactose. Pharm. Weekbl. (Sci.), 7 (1985) 186-193. Van der Voort Maarschalk, K., Zuurman, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., Porosity expansion of tablets as a result of bonding and deformation of particulate solids. Int. J. Pharm., 140 (1996) 185-193.

3 4

5 6

10

11

12

51

Chapter 3

Pore formation in tablets containing two materials

52

Matrix strength in tablets containing two materials

Chapter 4

Chapter 4 The influence of particles of a minor component on the matrix strength of sodium chloride

Summary This chapter deals with the matrix strength of sodium chloride particles in pure sodium chloride tablets and in tablets compressed from binary mixtures of sodium chloride with low concentrations pregelatinised starch. Because this study concerns the strength of the sodium chloride matrix, the tablet strength is reflected as a function of the sodium chloride volume fraction in the tablet. Starch particles in the mixture tablets decrease the sodium chloride volume fraction-tensile strength relation compared with that of pure sodium chloride tablets. To determine the contribution of the sodium chloride matrix to the tablet strength, the starch particles are removed from the mixture tablets by a heat treatment. Determination of the strengths of these heattreated tablets reveals that the sodium chloride matrix strength determines the tablet strength of mixture tablets containing a single matrix of sodium chloride particles. The decrease of the sodium chloride matrix density in the three different tablets (pure sodium chloride tablets, mixture tablets and heat-treated tablets) is reflected by an increase of the median pore size. The matrix in sodium chloride tablets shows a higher tensile strength to median pore size relation than the matrices in the mixture and heattreated tablets. Calculations assuming elastic-brittle fracture suggest that the initial presence of starch particles during tablet compaction causes the pores in the matrices of the mixtures and heat-treated tablets to be relatively flat and long. These pores weaken the sodium chloride matrix in the mixture and heat-treated tablets to a larger extent than the shorter more spherical pores formed during compaction of pure sodium chloride. Keywords: binary mixtures; tablet strength; matrix; median pore size; pore shape.

Published in Eur. J. Pharm. Sci. 16 (2002) 229-253.

53

Chapter 4

Matrix strength in tablets containing two materials

4.1

Introduction

The tensile strength of a tablet is often considered to be the result of the bonding between particles, which are the building blocks of tablets. Another property that is of relevance for the physical properties of the final tablet is the structure of the tablet. The structure is the resultant of the morphology of the compressed particles, which, together with the interactions between these particles, leads to a coherent tablet. In a tablet compressed from a mixture of two materials, three different types of interparticle bonding are found: cohesive forces between particles of the first component, cohesive forces between particles of the second component and the adhesive forces between particles of the different components. A change in the ratio of particles of the components in a binary mixture consequently results into a change in interparticle bonding and into a change of the structure in the tablet. The percolation theory, introduced by Stauffer (1) and Leuenberger (2), is commonly used to describe the particle structure in tablets compressed from binary mixtures. Particles of a material form a continuous matrix or exist in finite clusters. The transformation from finite clusters to a continuous matrix at a certain concentration in the tablet is defined as the percolation threshold. The presence of a continuous matrix is of crucial importance for the final strength of a tablet (3). Tablet compression and relaxation always results in the formation of pores. Fracture mechanics shows that not only pore size, but also pore shape and type of fracture are important factors for the tensile strength (4, 5) and are often as important as the interparticle attraction (6). This is because each individual pore is able to act as a stress concentrator (7, 8, 9). Studies of tablet strength of binary mixtures have reported contradictory results as an effect of non-identical experimental set-ups (e.g. 10, 11). However, it is mostly found that tablets compressed from binary mixtures show lower strength than calculated from linear interpolation of the tablet strengths of the single components, provided that the porosity has a constant value (12, 13, Ch. 2). The decrease of the tablet strength as a function of the tablet composition is different for every mixture. Apparently, the alterations in either the interparticle attraction or in the total tablet structure are such that they virtually always result in a strength decrease. The objective of this chapter is to study (1) the relation between the structure of a matrix and its strength and (2) the effects of the presence of another, minor, component on the tablet structure. Sodium chloride with a small fraction of air is the material that forms the structure, while the minor compound is pregelatinised starch.

54

Matrix strength in tablets containing two materials

Chapter 4

4.2 4.2.1

Materials and Methods Materials and mixtures

The materials used in this study were sodium chloride (glidant-free material, Akzo Nobel, Hengelo, The Netherlands) and pregelatinised potato starch (Paselli WA4, Avebe, Foxhol, The Netherlands). Because the sodium chloride batch contained coarse particles, the sodium chloride powder was milled (Moulinex, Birmingham, UK). The sodium chloride and pregelatinised starch samples were sieved with an Alpine Air Jet Sieve (Alpine, Augsburg, Germany) using USA Standard testing sieves (W.B. Tyler Inc., Mentor, OH, USA) in order to obtain the particle fraction between 106 and 212 m. Prior to use or measurements, the powders were stored at 20C and 60% RH for at least 4 days. The true densities of the conditioned 106-212 m fractions were measured at a temperature of 20C by helium pycnometry (Quantachrome, Syosset, NY, USA). The true densities of the 106-212 m fraction of sodium chloride was 2239 kgm-3 and of pregelatinised starch 1486 kgm-3. The binary mixtures (2.5-80% v/v) were blended in a Turbula mixer (Model 2P, W.A. Bachofen, Basle, Switzerland) at 90 rpm for a period of 15 min. 4.2.2 Tablets

Flat-faced compacts of 500 mg and a diameter of 13 mm were prepared on a compaction simulator (ESH, Brierley Hill, UK) at a temperature of 20 C and a relative humidity of 60%. The applied compaction pressures varied between 10 and 350 MPa. The die and punches were lubricated by dusting with magnesium stearate. While the lower punch was stationary during compaction, the upper punch displacements were sine waves with different amplitudes to vary the applied compaction pressures. The average punch speed during compaction was 3 mms-1. After a constant ejection time of 10 s, the relaxation period was at least 14 hours in a controlled climate chamber (Heraeus, Hanau, Germany) at 20C and 60% RH. The final dimensions of the tablets were measured with an electronic micrometer (Mitutoyo, Tokyo, Japan) and the weights were determined with an analytical balance (Mettler-Toledo, Greifensee, Switzerland). Tablet strength was measured with a Schleuniger 6D strength tester (Dr. Schleuniger Productronic, Soloturn, Switzerland). The tensile strength of tablets was calculated according to Fell and Newton (14). The relation between the tensile strength and the tablet porosity was determined for both pure materials and for every binary mixture containing a different percentage

55

Chapter 4

Matrix strength in tablets containing two materials

pregelatinised starch. The porosity-tensile strength relation was fitted using the Ryshkewitch-Duckworth relation (15, 16) using data of at least 40 tablets with different porosities. Using this fit, a tensile strength of tablets with a porosity of 15% was calculated for each different blend. 4.2.3 Heat treatment

To determine the structure formed by sodium chloride particles inside the tablet structure, three tablets of each mixture with a porosity of 15% ( 0.5%) were put in a muffle furnace (Naber Industrieofenbau, Bremen, Germany). In tablets compressed from pregelatinised starch and sodium chloride, the difference between the selfignition temperature of starch at 300C and the melting temperature of sodium chloride at 800C (17) enables the starch particles to be burned out, leaving the sodium chloride particles behind inside the tablet structure. Details of the heat treatment are described in Ch. 3. Dimensions, masses and tablet strengths of the tablets after the burning experiment were determined as described in sec. 4.2.2. 4.2.4 Pore size distributions

The pore size distributions of tablets compressed from the single materials and the binary mixtures and of the heat-treated tablets were measured by mercury intrusion porosimetry (Micromeritics, Model Autopore 9220, Norcross, GA, USA). Three tablets with a porosity of 15% ( 0.5%) or three heat-treated tablets were used for the determination of the pore size distribution of each binary mixture. Three tablets with equal porosity were used to measure the pore size distributions of sodium chloride tablets at each different sodium chloride volume fraction. Conditions and pressure ranges used during mercury intrusion measurements are described in Ch. 3. 4.2.5 Disintegration

The disintegration times of tablets were tested with a disintegration test unit according to the USP 24 test <701> (Model ZT 42, Erweka, Heusenstamm, Germany). To determine the disintegration time of tablets compressed from each blend containing a different percentage pregelatinised starch, the disintegration time of 5 tablets with a porosity of 15% ( 0.5%) was measured: the mean value is presented. The tests were conducted in 1 L tap water maintained at 37C and with no discs in the individual test compartments.

56

Matrix strength in tablets containing two materials

Chapter 4

4.3 4.3.1

Results and Discussion Structure-strength relationships

The tensile strength of a tablet is the consequence of its structure created by consolidation and relaxation. A frequently used comparison between the tablet structure and its strength is the porosity-tensile strength relation. Fig. 4.1a depicts this relation for tablets compressed from pure sodium chloride and from binary mixtures of sodium chloride with low concentrations of pregelatinised starch, respectively. All profiles show the same shape, but are numerically different. The pure sodium chloride tablets consisting of only a matrix of sodium chloride particles give the highest strength. The addition of the minor component (starch) to sodium chloride decreases the strength. Larger amounts of the minor component weaken the structure of the tablets to a larger extent.
8 Tensile strength (MPa) 6 4 2 0 0 0.05 0.1 0.15 Porosity (-) 0.2 0.25 0.3

Fig. 4.1a. Tensile strength as a function of the tablet porosity of tablets compressed from () sodium chloride, () binary mixtures of sodium chloride and 2.5% pregelatinised starch and () binary mixtures of sodium chloride and 15% pregelatinised starch.

This study is performed to determine the effect of the presence of a minor component on the sodium chloride structure. Therefore, it has been decided to correlate the alterations in strength to the volume fraction of sodium chloride rather than to total relative density or to porosity. The sodium chloride volume fraction is calculated as the fraction of the whole tablet volume, including air.

57

Chapter 4

Matrix strength in tablets containing two materials

8 Tensile strength (MPa) 6 4 2 0 1 0.95 0.9 0.85 0.8 0.75 0.7 Sodium chloride volume fraction (-)

Fig. 4.1b. Tensile strength as a function of the tablets sodium chloride volume fraction. Symbols as in Fig. 4.1a.

Fig. 4.1b shows this sodium-chloride-volume-fraction to tensile strength relation of the tablets given in Fig. 4.1a. As described for the porosity-tensile strength relation, the sodium-chloride-volume-fraction to tensile strength relations show lower tensile strengths as a result of the addition of starch to the sodium chloride structure. Fig. 4.1b shows that the tensile strength becomes a function of the sodium chloride only when the sodium chloride volume fraction is higher than about 0.82. It is not possible to make a clear distinction between these three types of tablets. In contrast, when the sodium chloride volume fraction is lower than about 0.82, the tablets that were compressed in the presence of starch show equal tensile strengths. Apparently, starch has no concentration-dependent influence on the strength of the sodium chloride matrix in this density range. However, starch must have had an effect during compression, because the tablets that were compressed in the presence of starch are always weaker than the tablets compressed from pure sodium chloride at equal sodium chloride fraction. In order to study the effects of starch on the matrix strength of sodium chloride in more detail, the starch particles were removed from tablets compressed from mixtures containing low amounts of starch. The individual tablet structure formed by the sodium chloride particles was obtained by burning out the starch particles in an oven. Because sodium chloride is a material that always forms a tablet structure containing a fraction of air, the air fraction before removal of the starch particles was fixed at 0.15 for all tablets compressed from mixtures. Removal of increasing amounts of starch

58

Matrix strength in tablets containing two materials

Chapter 4

from the mixture tablets results in heat-treated tablets consisting of only the sodium chloride matrix and increasing air fractions. The sodium chloride volume fraction-tensile strength relation of these heattreated tablets is given in Fig. 4.2 and is compared with the relation of the pure sodium chloride tablets (same as depicted in Fig. 4.1b). The tensile strength of the heat-treated tablet with a sodium chloride volume fraction of 0.85, i.e. no starch particles were present in the tablet before the heat treatment, shows hardly any difference with that of the sodium chloride tablet with equal sodium chloride volume fraction. This indicates that the heat treatment has a negligible effect on the sodium chloride matrix and its strength. This is supported by the mercury intrusion porosimetry measurements of the pure sodium chloride tablets reported in Ch. 3.
6 Tensile strength (MPa) 5 4 3 2 1 0 0.85

0.64

0.43

0.21

0.00

Sodium chloride volume fraction (-)

Fig. 4.2. Tensile strength as a function of the sodium chloride volume fraction of () sodium chloride tablets, () heat-treated tablets and () tablets compressed from binary mixtures of sodium chloride and pregelatinised starch. Error bars are the 95% confidence intervals of the heat-treated tablets.

As can be observed from Fig. 4.2, the tensile strengths of the heat-treated tablets are systematically lower than the tensile strengths of the sodium chloride tablets. Apparently, the sodium chloride matrix in the heat-treated tablet has been weakened by the initial presence of starch particles. The heat-treated tablets with sodium chloride volume fractions of 0.60 and lower did not retain their original dimensions and showed no strength. Apparently, the sodium chloride particles cannot form a continuous matrix any longer in these tablets compressed from mixtures. For this reason, in tablets compressed from blends of sodium chloride and pregelatinised

59

Chapter 4

Matrix strength in tablets containing two materials

starch, a sodium chloride volume fraction between 0.64 and 0.60 can be considered as the percolation threshold of the sodium chloride matrix. The third type of tablets shown in Fig. 4.2 are the tablets compressed from the mixtures of sodium chloride and starch with an air fraction of 0.15 without the heat treatment; the starch particles are still present in the tablets. Tablets consisting of high volume fractions sodium chloride (and consequently low volume fractions starch) have the same decrease in tensile strength as the heat-treated tablets. The tablet strengths are comparable and are independent of the presence of the starch particles in the tablet structure. From this observation it can be concluded that the strength of the sodium chloride matrix entirely determines the tablet strength. The contribution of the adhesion forces of the starch particles to the matrix strength and the tablet strength can be considered as zero. In contrast, tablets compressed from mixtures with a sodium chloride volume fraction lower than 0.76 show higher tensile strengths than the heat-treated tablets with the same initial composition. Obviously, the tensile strength of these tablets is not determined anymore by the strength of sodium chloride matrix. A change in tablet structure terminates the domination of the matrix strength of sodium chloride particles. A disintegration test of tablets can support this suggestion. Luginbhl and Leuenberger (18) showed that the minimum disintegration time for tablets containing starch as a disintegrant is found for tablets in which starch particles start to form a matrix. Table 4.1 lists the disintegration times of tablet compressed from the mixture of sodium chloride and pregelatinised starch with an air fraction of 0.15.
Table 4.1. Disintegration times of tablets compressed from binary mixtures of sodium chloride and pregelatinised starch with an air fraction of 0.15 Sodium chloride volume fraction (-) 0.85 0.83 0.81 0.79 0.76 0.72 0.68 0.64 0.60 0.51 Average disintegration time (s). Standard deviation in parentheses. 206 (7) 200 (11) 99 (13) 69 (10) 34 (3) 49 (5) 75 (11) 323 (64) 1930 (200) 5102 (250)

60

Matrix strength in tablets containing two materials

Chapter 4

The table shows the shortest disintegration times for tablets with a sodium chloride volume fraction of 0.76. Both the end of the influence of the sodium chloride matrix on the tablet strength (Fig. 4.2) and the disintegration test prove that the percolation threshold of the starch matrix is found in tablets with a sodium chloride volume fraction of 0.76. Chapter 3 of this thesis states that besides the changes in particle structure, the pore size distributions in tablets compressed from mixtures also change as a function of the tablet composition. Because of the pores ability to concentrate stress, pore characteristics are very important for changes in tablet and matrix strength. 4.3.2 Pore size-matrix strength relationships

Tablets consist of a fraction solids and a complementary fraction of air (porosity). The air fraction is built up from pores with different sizes and is represented by the pore size distribution.
0.01 Pore size distribution (ml m-1 g-1) 0.008 0.006 0.004 0.002 0 100

10 Pore diameter (m)

Fig. 4.3. Pore volume size distributions of sodium chloride tablets with sodium chloride volume fractions of () 0.85, () 0.83 and () 0.79.

In Fig. 4.3, the pore size distributions measured by mercury intrusion porosimetry are given for sodium chloride tablets containing different volume fractions of sodium chloride. Upon decreasing sodium chloride volume fractions, the pore size distributions of tablets show a shift to larger pores while the distribution of the small pore sizes (< 5 m) hardly changes in size and number. The pore size distributions of tablets compressed from mixtures of sodium chloride and pregelatinised starch also show a change in the larger pores, while the smaller pores stay constant in size and

61

Chapter 4

Matrix strength in tablets containing two materials

number (Ch. 3). To express this change in pore size distribution in a single value, the median pore diameter was calculated for all tablets containing a sodium chloride matrix. Fig. 4.4a shows the sodium chloride volume fraction-median pore diameter relation of the pure sodium chloride tablets, tablets compressed from the mixtures and heat-treated tablets, all containing a single matrix formed by sodium chloride particles. Tablets containing also a matrix of starch particles are not shown in Fig. 4.4a in comparison with Fig. 4.2. Every type of these tablets shows an increase of the median pore diameter as an effect of the decreased sodium chloride volume fraction. The sodium chloride tablets have higher median pore diameters than the tablets compressed from the mixtures and the heat-treated tablets, proving the presence of larger pores in the sodium chloride matrix. Although the heat-treated tablets and the tablets compressed from the mixtures have identical sodium chloride matrices, the heat-treated tablets show a higher median pore diameter than the tablets compressed from the mixtures. Apparently, the removed starch particles are slightly larger than the original pores, causing an increase of the median pore diameter. This is confirmed by Ch. 3.
Median pore diameter ( m) 26 22 18 14 10 6 0.85

0.8

0.75

0.7

Sodium chloride volume fraction (-)

Fig. 4.4a. Median pore diameters measured by mercury intrusion porosimetry as a function of the sodium chloride volume fraction. Symbols as in Fig. 4.2.

The theory of fracture mechanics describes that the combination of size, orientation and shape of the pores determines their influence on the tensile strength of the tablet. In general, pores with the largest length perpendicular to the stress profile in the tensile strength measurement cause the failure of the material, as these pores give the highest stress concentration. Unfortunately, the exact pore dimensions (length and

62

Matrix strength in tablets containing two materials

Chapter 4

diameter) cannot be determined by mercury intrusion porosimetry. This technique measures the opening to the pore volume; thereby it underestimates in many cases the actual pore dimensions. However, closer packing of the particles in the tablet will decrease the pore opening as well as the pore dimensions (19). As a consequence, a decrease of the median pore diameter mainly reflects a decrease of the dimensions of the larger pores. The relation between the median pore diameter and the strength of the tablets depicted in Fig. 4.4a is shown in Fig. 4.4b. The decreasing effect of larger median pore diameters on the sodium chloride matrix strength is valid for all three types of tablets. This phenomenon has also been reported for compacts compressed from other materials (20) and from ceramic materials (21). The heat-treated and mixture tablets show a similar relation between the median pore diameter and the tensile strength. These two types of tablets have the same sodium chloride matrices that determine the tablet strengths (Fig. 4.2). On the contrary, the sodium chloride tablets show higher tensile strengths at comparable median pore diameters. This indicates that the initial presence of starch particles during compaction of the binary powder mixtures affected the influence of the median pore size on the sodium chloride matrix strength.
4 Tensile strength (MPa) 3 2 1 0 6 10 14 18 22 26 Median pore diameter (m)

Fig. 4.4b. Tensile strength of sodium chloride matrix as a function of the median pore diameter. Symbols as in Fig. 4.2.

63

Chapter 4

Matrix strength in tablets containing two materials

4.3.3

Pore shape in sodium chloride matrix

The effective influence of the median pore diameter on the sodium chloride matrix strength was found to be different for the sodium chloride matrices in the different tablets (Fig. 4.4b). The effective influence on the tensile strength is not only dependent on the size, but also on other factors, like pore shape. It would be interesting to calculate the effective pore dimensions in the different sodium chloride matrices and compare the results with Scanning Electron Microscopy photographs. However, as already mentioned, mercury intrusion data are interpreted as round pore openings based on mercury intrusion pressures. The median pore diameter gives an interpreted size of the large pores rather than the real large pore dimensions, i.e. length and diameter. Implementation of the median pore diameter as the large pore dimension results in unreliable and imprecise effective pore dimensions. However, the median pore diameter still is a useful indication of the size of the large pores. To give an estimation of the pore shapes in the individual sodium chloride matrices, the pore shapes were calculated according to the Inglis-Griffith theory of elastic-brittle fracture (4, 22). Although sodium chloride and starch are known as plastically deforming materials, compacted powders behave brittle under tensile loading (23). This means that the tablet fracture is not preceded by any significant plastic deformation. The relation between the pore shape and decrease of the material strength derived by Inglis and Griffith is reflected by the equation:

max measured

= 1+ 2

(1)

where max is the theoretical maximum tensile strength, measured is the measured tensile strength, l is the total length of the pore and is the diameter of the curvature of the pore tip. The pores are taken to be as ellipsoidal holes with the mathematical equation:

d2 l

(2)

in which d is the diameter of the pore. The advantage of taking pores to be ellipsoidal holes is that the pore shape can vary from a spherical hole (l = d) to a sharp crack (l >> d).

64

Matrix strength in tablets containing two materials

Chapter 4

By rewriting eq. 1, it is possible to describe the pore shape, expressed by the length-todiameter ratio of the pore, l/d:

l max = 1 2 d measured

(3)

To be able to calculate the pore shapes in the different sodium chloride matrices, max has to be determined. The theoretical maximum tensile strength max is the failure of atomic bonds across an atomic plane. A fracture plane will arise when the stress at the tip of the pore reaches max. The theoretically correct value of max would be the tensile strength of a single sodium chloride crystal. Since this value is hard to obtain, the tensile strength at zero tablet air fraction from the Ryskewitch-Duckworth equation of the sodium chloride tablets (max = 11.4 MPa, r2 = 0.976) is considered to provide a good approximation of the maximum theoretical tensile strength. This single value of the maximum theoretical tensile strength was used as maximum tablet strength for all tablets containing the different sodium chloride matrices. Using this information, the shapes of the pores have been calculated as a function of the sodium chloride volume fraction. Fig. 4.5 gives these pore shapes of the sodium chloride, heat-treated and mixture tablets, all containing a single sodium chloride matrix. The mixture and heat-treated tablets show higher values of the pore shape than the sodium chloride tablets. This suggests that the pores in the first-mentioned tablets are relatively flatter and/or have sharper tips at the endpoints (eq. 2).
10 8 l/d (-) 6 4 2 0 0.85

0.8

0.75

0.7

Sodium chloride volume fraction (-)

Fig. 4.5. Pore shape expressed as l/d-ratio as a function of the sodium chloride volume fraction in tablets containing a single matrix of sodium chloride particles. Symbols as in Fig. 4.2.

65

Chapter 4

Matrix strength in tablets containing two materials

The development of these pores can be explained by examining the compaction behaviours of mixtures. As a result of the difference in yield pressure, the softer starch particles are flattened between the rigid sodium chloride particles during the densification of the mixtures (Ch. 2). Relaxation of tablets compressed from these mixtures results in larger pores (Ch. 3), which can be considered to adopt the shape of relatively flat holes. It is expected that these relatively flat pores weaken the sodium chloride matrix to a larger extent than the more spherical shaped pores (eq. 3) formed in tablets compressed from sodium chloride only. The consequence of this finding is that the extent of particle deformation of softer materials is of importance on the pore shape in the main structure formed by more rigid materials. This difference in rigidity (difference in yield pressure) influences the matrix formation (24) as well as the creation of the pore shape.

4.4
1 2 3

References
Stauffer, D., Introduction to percolation theory. Taylor & Francis, London (1985) p. 3103. Leuenberger, H., Compression of binary powder mixtures and solubility parameters of solids. Int. J. Pharm., 27 (1985) 127-138. Holman, L.E., Leuenberger, H., The effect of varying the composition of binary powder mixtures and compacts on their properties: A percolation phenomenon. Powder Techn., 60 (1990) 249-258. Griffith, A.A., Phenomena of rupture and flow in solids. Phil. Trans. Roy. Soc., A221 (1920) 163-168. Griffith, A.A., Theory of rupture. Proc. 1st Int. Conf. Appl. Mech., J. Waltman Jr., Delft (1924) 55. Al-Nasassrah, M., Podczeck, F., Newton, J.M., The effect on an increase in chain length on the mechanical properties of polyethylene glycols. Eur. J. Pharm. Biopharm., 46 (1998) 31-38. Rowe, R.C., Roberts, R.J., Simulation of crack propagation in porous compacted specimens of microcrystalline cellulose. Eur. J. Pharm. Biopharm., 40(1) (1994) 9-13. Hiestand, E.N., Mechanical properties of compacts and particles that control tableting success. J. Pharm. Sci., 86(9) (1997) 985-990. Hiestand, E.N., Principles, tenets and notions of tablet bonding and measurements of strength. Eur. J. Pharm. Biopharm., 44 (1997) 229-242. Cook, G.D., Summers, M.P., Effect of compression speed on the tensile strength of tablets of binary mixtures containing aspirin. J. Pharm. Pharmacol., 42 (1990) 462-467.

4 5 6

7 8 9 10

66

Matrix strength in tablets containing two materials 11

Chapter 4

Mattsson, S., Nystrm, C., Evaluation of strength-enhancing factors of a ductile binder in direct compression of sodium bicarbonate and calcium carbonate powders. Eur. J. Pharm. Sci., 10 (2000) 53-66. Sheikh-Salem, M., Fell, J.T., Compaction characteristics of mixtures of materials with dissimilar compaction mechanisms. Int. J. Pharm. Tech. Prod. Mfr., 2(1) (1981) 19-22. Blattner, D., Kolb, M., Leuenberger, H., Percolation theory and Compactibility of binary powder systems. Pharm. Res., 7(2) (1990) 113-117. Fell, J.T., Newton, J.M., The tensile strength of lactose tablets. J. Pharm. Pharmacol., 20 (1968) 657-658. Ryshkewitch, E., Compression strength of porous sintered alumina and zirconia. J. Am. Ceram. Soc., 36 (1953) 65-68. Duckworth, W.H., Discussion of Ryshkewitch paper by Winston Duckworth. J. Am. Ceram. Soc., 36 (1953) 68. Weast, R.C., Astle, M.J., CRC Handbook of chemistry and physics, 60th edition. CRC Press Inc., Boca Raton, Florida (1979-1980). Luginbhl, R., Leuenberger, H., Use of percolation theory to interpret water uptake, disintegration time and intrinsic dissolution rate of tablets consisting of binary mixtures. Pharm. Acta Helv., 69 (1994) 127-134. Rice, R.W., Porosity effects on machining direction strength anisotropy and failure mechanisms. J. Am. Ceram. Soc., 77 (1994) 2232-2236. Juppo, A.M., Relationship between breaking force and pore structure of lactose, glucose and mannitol tablets. Int. J. Pharm., 127 (1996) 95-102. Liu, D-M., Influence of porosity and pore size on the compressive strength of porous hydroxyapatite ceramic. Ceram. Int., 23 (1997) 135-139. Inglis, C.E., Stresses in a plate due to the presence of cracks and sharp corners. Roy. Inst. Naval Architects Trans., 55 (1913) 219-230. Stanley, P., Mechanical strength testing of compacted powders. Int. J. Pharm., 227 (2001) 27-38. Kuentz, M., Leuenberger, H., A new theoretical approach to tablet strength of a binary mixture consisting of a well and a poorly compactable substance. Eur. J. Pharm. Biopharm., 49 (2000) 151-159.

12 13 14 15 16 17 18

19 20 21 22 23 24

67

Chapter 4

Matrix strength in tablets containing two materials

68

Effect of particle size on matrix strength

Chapter 5

Chapter 5 The mutual effects of initial particle size and addition of another compound on tablet strength

Summary The purpose of this chapter is to describe the effect of particle size on the matrix strength in tablets compressed from pure sodium chloride and from blends containing starch as minor component. It has been shown in the previous chapter that the removal of the starch particles by a heat treatment made it possible to measure the percolation thresholds and strengths of the sodium chloride matrices in the tablet structure. The percolation threshold of sodium chloride is directly related to its mean particle size, whereas the percolation threshold of starch hardly changes at low starch volume fractions. Starch particles cause an increase in median pore diameter in tablets compressed from blends. Alterations in tensile strength of the tablets compressed from blends are related to the median pore diameter. A unique relation has been found between the median pore diameter and the tensile strength of the tablets prepared from mixtures. This relation is independent of the sodium chloride mean particle size and sodium chloride volume fraction. A similar relation exists for tablets compressed from only sodium chloride particles, but higher tensile strengths were observed. The deviation between both relations is explained by a difference in pore shape as defined in Ch. 4. The pore shapes in the tablets compressed from blends are flattened in contrast to the more spherical pores in the pure sodium chloride tablets. Keywords: particle size distribution; tablet strength; binary mixtures; median pore size; pore shape.

Submitted for publication in Int. J. Pharm.

69

Chapter 5

Effect of particle size on matrix strength

5.1

Introduction

It is widely accepted that the strength of compacts like tablets, is the result of bonding between particles. This is, for example, expressed by Rumpfs classical equation that correlates the strength of a body with its porosity, the size of the building blocks and the attractive force between the particles (1). However, from fracture mechanics it is also known that pores and other discontinuities also affect the mechanical properties of bodies (2, 3, 4). For compacts of single materials it was found that the presence of pores with a certain size and shape affects the strength of compacts as much as the interparticle bonding does (5). This paper expands on the results given in Ch. 4, which discusses the negative effect of the addition of small amounts of a second compound (starch) on the strength of the matrix of a major compound (sodium chloride). It shows that the addition of the minor compound with a lower Youngs modulus causes formation of pore structures with anisomorphous shapes, i.e. the pores are flat and have sharp edges. The resulting pore shape explains why the porosity effect in tablets prepared from blends is different from that of tablets compressed from single materials. Clearly, it is not only the reduction in interparticle bonding that induces the reduction in strength of tablets made from (formulated) powder blends, which will be further described in Ch. 6, but there is also another pore related parameter influencing matrix strength. The discussion so far does not consider the effect of a change in particle size distribution (of one of the components) on matrix structure and resulting tablet strength, while this can be of great relevance for formulation practice. A study performed by Caraballo et al. (6) already showed that materials with a larger particle size need a higher concentration in the powder blend to percolate the tablet structure. Although the influences of particle size on percolation threshold are generally understood, the changes in percolation thresholds by interactions between the different materials in a powder blend are currently unknown. This chapter aims to relate the effect of sodium chloride particle size to the matrix structure and strength.

70

Effect of particle size on matrix strength

Chapter 5

5.2 5.2.1

Materials and Methods Materials and mixtures

The materials used were sodium chloride (glidant-free material, Akzo Nobel, Hengelo, The Netherlands) and pregelatinised potato starch (Paselli WA4, Avebe, Foxhol, The Netherlands). Prior to sieving, the sodium chloride powder was milled (Moulinex, Birmingham, UK). The sodium chloride and pregelatinised starch powders were sieved with an Alpine Air Jet Sieve (Alpine, Augsburg, Germany) equipped with USA Standard testing sieves (W.B. Tyler Inc., Mentor, OH, USA) in order to obtain the desired particle size fractions. The sieve fractions of sodium chloride used were 0-38 m, 53-106 m and 106-212 m. The 106-212 m particle size fraction of starch has been used. After sieving, the powders were stored at 20 C and 60% RH for at least 4 days. The volume-based mean particle sizes of the particle size distributions were obtained by laser diffraction (Sympatec, Helos KA/LA, Clausthal-Zellerfeld, Germany). Table 5.1 lists the mean particle sizes of the powders used in this paper. The true densities of the powders were measured at a temperature of 20C by helium pycnometry (Quantachrome, Syosset, NY, USA). The true densities of sodium chloride and pregelatinised starch were 2239 kgm-3 and 1486 kgm-3, respectively. The binary mixtures were mixed in a Turbula mixer (Model 2P, W.A. Bachofen, Basle, Switzerland) at 90 rpm for a period of 15 min.
Table 5.1: Powder characteristics of sodium chloride and pregelatinised starch. Material Particle size Mean particle Percolation threshold Percolation threshold starch1 sodium chloride1 range (m) size (m) Sodium chloride 0-38 14 0.35 0.68 53-106 83 0.51 0.76 106-212 170 0.60 0.76 Starch 106-212 175 1 Percolation thresholds are given as a function of the sodium chloride volume fraction of tablets containing a porosity of 0.15.

5.2.2

Tablets

Tablets (flat-faced, 500 mg, diameter 13 mm) were compacted on a compaction simulator (ESH, Brierley Hill, UK) at a temperature of 20C and a relative humidity of 60%. Before each compaction cycle, the die and punches were lubricated with

71

Chapter 5

Effect of particle size on matrix strength

magnesium stearate using a brush. The upper punch displacements were sine waves with different amplitudes in order to reach the applied compaction pressures. The average upper punch speed was 3 mms-1. The lower punch was stationary during compaction, and the ejection time was always 10 s. The relaxation period after compaction was at least 14 hours in a controlled climate chamber (Heraeus, Hanau, Germany) at 20C and 60% RH. The final dimensions of the tablets were measured with an electronic micrometer (Mitutoyo, Tokyo, Japan) and weighed with an analytical balance (Mettler-Toledo, Greifensee, Switzerland). Tablet strength was measured with a Schleuniger 6D strength tester (Dr. Schleuniger Productronic, Soloturn, Switzerland). The tensile strength of tablets was calculated according to Fell and Newton (7). The tensile strengths and the porosities of the tablets compressed from the individual materials and each binary mixture were related to each other and fitted by the Ryskewitch-Duckworth equation (8, 9) using data of at least 40 tablets with different porosities. Using this fit, the tensile strength at a porosity of 15% was calculated for tablets compressed from the single materials containing the different particle size fractions and the different blends. 5.2.3 Heat treatment

To examine the structure and strength of matrices formed by sodium chloride inside the tablet structures made from blends, five tablets of each mixture with a porosity of 15% ( 0.5%) were put in a muffle furnace (Naber Industrieofenbau, Bremen, Germany). Details of the heat treatment are described in Ch. 3. It was possible to remove at least 85% of the total amount of starch present in the tablets. Dimensions, weights and crushing strengths of the tablets after the burning experiment were determined as described in sec. 5.2.2. 5.2.4 Pore size distributions

The pore size distributions of tablets compressed from the single materials and the binary mixtures were measured by mercury intrusion porosimetry (Micromeritics, Model Autopore 9220, Norcross, GA, USA). Three tablets with a porosity of 15% ( 0.5%) were used for the measurement of the pore size distribution of each blend. Details of the mercury intrusion porosimetry measurements are described in Ch. 3. The surface tension of mercury used in the calculations was 480 mNm-1 and the contact angle of mercury with sodium chloride and starch was assumed to be 140 in all situations.

72

Effect of particle size on matrix strength

Chapter 5

5.3 5.3.1

Results and Discussion Starch affecting sodium chloride matrix strength

Fig. 5.1 gives the tensile strength-porosity relations of tablets compressed from sodium chloride with different particle sizes. The figure shows that all strengthporosity profiles have similar shapes, but the profiles are numerically different. This is consistent with other observations (10).
10 Tensile strength (MPa) 8 6 4 2 0 0 0.1 0.2 Porosity (-) 0.3 0.4

Fig. 5.1. Tensile strength as a function of the porosity of tablets compressed from sodium chloride with particle sizes between () 0-38 m, () 53-106 m and ([) 106-212 m, respectively.

Ch. 4 already has shown that the addition of a small amount of pregelatinised starch to sodium chloride powder decreases the tensile strengths of tablets compressed from mixtures and reduces the strength of the sodium chloride matrix too. In order to test the strengths of sodium chloride matrices consisting of different particle sizes in tablets compressed from blends, tablets were prepared with a constant air fraction (porosity) of 0.15. Fig. 5.2a-5.2c show the tensile strengths of these tablets as a function of the sodium chloride volume fraction before and after the removal of the starch particles by a heat treatment. Up to a certain sodium chloride volume fraction, the tensile strengths of the heat-treated tablets are comparable to the initial tensile strengths of the tablets compressed from the blends. According to Ch. 4, these tablet structures consist of sodium chloride matrices combined with clusters of starch particles. The sodium chloride matrices completely determine the tensile strength of tablets compressed from mixtures when the starch particles exist as clusters.

73

Chapter 5

Effect of particle size on matrix strength

7 6 5 4 3 2 1 0 0.85

Tensile strength (MPa)

0.64

0.43

0.21

0.00

Sodium chloride volume fraction (-)

Tensile strength (MPa)

6 5 4 3 2 1 0 0.85 0.64 0.43 0.21 0.00

Sodium chloride volume fraction (-)

Tensile strength (MPa)

6 5 4 3 2 1 0 0.85 0.64 0.43 0.21 0.00

Sodium chloride volume fraction (-)

Fig. 5.2. Tensile strength of tablets with an initial air fraction of 0.15 before (open symbols) and after heat treatment (closed symbols). Tablets are compressed from binary mixtures of pregelatinised starch (106-212 m) and sodium chloride with particle size distribution of: (a) 0-38m (), (b) 53-106 m () and (c) 106-212 m ([). Error bars are the 95% confidence intervals of the tensile strength after the heat treatment.

74

Effect of particle size on matrix strength

Chapter 5

At sodium chloride volume fractions lower than the sodium chloride volume fractions of about 0.68 (0-38 m sodium chloride) or 0.76 (53-106 m sodium chloride and 106-212 m sodium chloride), the heat-treated tablets have lower tensile strengths than the non-treated tablets. These sodium chloride volume fractions reflect the percolation thresholds of the starch matrix and are listed in table 5.1. Heat-treated tablets containing sodium chloride volume fractions lower than approximately 0.35 (038 m sodium chloride), 0.51 (53-106 m sodium chloride) or 0.60 (106-212 m sodium chloride) partly disintegrated during the heat treatment and it was impossible to measure any tablet strength. Apparently, the sodium chloride particles exist as clusters instead of a continuous network inside these tablet structures. These sodium chloride volume fractions reflect the percolation thresholds of sodium chloride consisting of the different particle size fractions (Table 5.1). As different particle sizes of sodium chloride are combined with starch with a constant mean particle size, the arrangement of the particles in the tablet structure changes. The percolation thresholds of sodium chloride and starch are represented in Fig. 5.3 as a function of the sodium chloride mean particle size.
Percolation threshold (sodium chloride volume fraction) 0.8 0.7 0.6 0.5 0.4 0.3 0 50 100 150 200 Sodium chloride mean particle size (m)

Fig. 5.3. Percolation threshold of () the sodium chloride matrix and () the pregelatinised starch matrix, respectively, as a function of the sodium chloride mean particle size in mixture tablets with an air fraction of 0.15.

Fig. 5.3 shows that the percolation thresholds of the sodium chloride matrices are found at lower sodium chloride volume fractions when the particle size decreases. The percolation threshold of sodium chloride in tablets compressed from the blends is proportional to its particle size. On the one hand, smaller particle sizes enable sodium chloride to establish a matrix at lower volume fractions. On the other hand, the

75

Chapter 5

Effect of particle size on matrix strength

percolation thresholds of starch hardly change with sodium chloride mean particle size variations. The percolation threshold of starch remains unchanged, as a constant particle size distribution of starch has been used for all different blends. Apparently, there is only a limited effect of the sodium chloride particle size on the percolation threshold of starch. It is remarkable that the starch particles already form a matrix at low starch fractions, even when the initial particle size of sodium chloride is much smaller than that of starch. Due to the difference in yield pressure between sodium chloride and starch, the starch particles are highly deformed between the rigid sodium chloride particles during compaction (Ch. 3). Therefore, the starch particles will already establish a percolating network throughout the structure at low starch fractions. In contrast, the tensile strength of tablets compressed from blends is clearly influenced by the mean particle size of sodium chloride (Fig. 5.2a-5.2c). The tablets containing small sodium chloride particles have higher tensile strengths and a lower reduction in tensile strength than tablets containing large sodium chloride particles. To describe the negative influences of starch on the sodium chloride matrix strengths, changes in matrix structures are further characterised. 5.3.2 Effect of pore size on matrix strength

Tablets compressed from sodium chloride not only consist of particles with different particle sizes, but also of pores between the sodium chloride particles. Since pores have the ability to concentrate stress, they influence the tensile strength of a porous compact. In general, pores with the largest diameters perpendicular to the applied stress cause the failure of a porous compact (3, 4). Furthermore, Ch. 4 states that the median pore size adequately describes the degree of strength reduction of sodium chloride compacts. Fig. 5.4a shows the median pore diameters of the tablets compressed from mixtures with an air fraction of 0.15. This figure only depicts the median pore diameter of the tablets of which the sodium chloride matrix completely determines the tablet strength. An increase in median pore diameter upon decreasing sodium chloride volume fractions is found for the different sodium chloride particle sizes. The value of the median pore diameter depends on the particle size of sodium chloride. A larger sodium chloride particle size results in higher median pore diameters in both the pure sodium chloride tablets and the tablets compressed from blends. Moreover, the increase in median pore diameter is also higher when sodium chloride particles in the tablets are larger. The starch particles influence the increase of

76

Effect of particle size on matrix strength

Chapter 5

the median pore diameter to a greater extent when the sodium chloride particles are coarser.
Median pore diameter ( m) 12 10 8 6 4 2 0 0.85

0.8

0.75

0.7

0.65

Sodium chloride volume fraction (-)

Fig. 5.4a. Median pore diameter measured by mercury intrusion porosimetry as a function of the sodium chloride volume fractions in tablets compressed from binary mixtures of pregelatinised starch (106-212 m) and sodium chloride with particle sizes between () 0-38 m, () 53-106 m and ([) 106-212 m, respectively. The arrows point to the percolation thresholds of pregelatinised starch.

In Fig. 5.4b, the median pore diameter of the different tablets is plotted against the tensile strength.
6 Tensile strength (MPa) 5 4 3 2 1 0 0 2 4 6 8 10 12 Median pore diameter (m)

Fig. 5.4b. Tensile strength of tablets compressed from () pure sodium chloride with a sodium chloride volume fraction of 0.85 and of the mixture tablets (symbols as in Fig. 5.4a.) as a function of the median pore diameter.

77

Chapter 5

Effect of particle size on matrix strength

Fig. 5.4b shows that the median pore size is an important parameter to describe the tensile strength of pure sodium chloride tablets with equal porosities of 0.15. Other authors (11, 12) also stated that the median pore size is important for the characterisation of the compact strength. In analogy with Ch. 4, there exists a relationship between the median pore diameter and the tensile strengths of tablets compressed from blends (Fig. 5.4b). The median pore diameter appears to be the controlling parameter of the tensile strength of the tablets compressed from mixtures, irrespective of the different sodium chloride fractions and sodium chloride mean particle sizes. Therefore, it can be concluded that the reduction in tensile strength by the addition of starch is mainly the result of an increase in median pore diameter (Fig. 5.4a). However, the tablets consisting of blends have lower tablet strengths than the pure sodium chloride tablets at corresponding median pore diameters (Fig. 5.4b). The initial presence of starch during compaction affects the influence of the median pore diameter on the sodium chloride matrix strength. This indicates that there must be another factor besides the median pore diameter that determines the tablet strength. 5.3.3 Pore shape in relation to median pore size

The theory of fracture mechanics implies that not only the size of pores, but also the pore shape is important for the tensile strength of a porous compact. In Ch. 4, the pore shape is defined as the length-to-diameter ratio, l/d, and can be correlated with the tensile strength:

max 1 l measured = d 2

(1)

where l is the total pore length, d the diameter of the pore, max the maximum material strength and measured the measured tensile strength. The curvature at the tip of the pore is given by d2/l. To calculate the pore shape in the different sodium chloride matrices, it is necessary to obtain an accurate value of max. The correct maximum strength is the tensile strength of a sodium chloride crystal with the dimensions of a tablet without any fault, cracks or pores. The tensile strength at zero porosity is considered to be a good approximation of the maximum tensile strength. The extrapolation of the three relationships (0-38 m, 53-106 m and 106-212 m sodium chloride) given in Fig. 5.1a by the Ryskewitch-Duckworth equation results into tensile strengths at zero air fraction of 14.9 MPa, 13.4 MPa and 11.4 MPa, respectively. Where the tensile

78

Effect of particle size on matrix strength

Chapter 5

strength at zero air fraction increases with decreasing particle size, the tablet strength at zero air fraction will increasingly be affected by the particle bonding than by the pore size. Therefore the best value of max that can be obtained by compaction has been calculated by linear extrapolation to infinitely small sodium chloride particles (max = 15.2 MPa, r2 = 0.999). This single value is used as max in further calculations of the pore shapes.
8 6 l/d (-) 4 2 0 0.85

0.8

0.75

0.7

0.65

Sodium chloride volume fraction (-)

Fig. 5.5. Pore shape defined as the l/d-ratio in mixture tablets compressed from blends of pregelatinised starch (106-212 m) and sodium chloride with particle size fractions of () 0-38 m, () 53-106 m and ([) 106-212 m, respectively.

The pore shapes in the sodium chloride tablets and the mixture tablets of which the sodium chloride matrix controls the tablet strength are calculated by eq. 1 and given in Fig. 5.5. The l/d-ratios of the pore shapes in the pure sodium chloride tablets (sodium chloride volume fraction of 0.85) have low values. Generally, the addition of starch to the sodium chloride matrix increases the l/d-ratio, indicating that the pores are relatively longer and/or have sharper tips. The effect of the starch addition on pore shape is relatively small for the tablets containing the 0-38 m sodium chloride particles. Similarly as has been described for the relationship between median pore size and sodium chloride volume fractions, it seems that the influence of the starch particles on the pore formation in the matrix structure depends on the sodium chloride particle size. The compaction of starch particles in combination with fine sodium chloride particles leads to a relatively less anisomorphous pore shapes (l/d-ratios are approximately 1) in contrast to the high l/d-ratios found when larger sodium chloride particles are used. This difference in pore shape is likely to be caused by local phenomena in the tablet structure occurring during tablet relaxation.

79

Chapter 5

Effect of particle size on matrix strength

The calculated pore shapes and median pore diameters of the sodium chloride matrices are combined in Fig. 5.6. The tablets compressed from blends have a slightly higher l/d-ratio at low to medium median pore diameters compared to those of the pure sodium chloride tablets. At larger median pore diameters, this difference in pore shape is larger, i.e. the larger pores have flatter shapes and sharper tips. This in turn explains the lower median pore diameter-tensile strength relation of tablets compressed from blends found in Fig. 5.4b.
8 6 l/d (-) 4 2 0 0 2 4 6 8 10 12 Median pore diameter (m)

Fig. 5.6. Pore shape in tablets compressed from () pure sodium chloride with a sodium chloride volume fraction of 0.85 and in the mixture tablets (symbols as in Fig. 5.5.) as a function of the median pore diameter.

In conclusion, the sodium chloride matrix strength and the effect of a minor component on the matrix strength can be characterised by a combination of changes in median pore size and in pore shape, where shape seems to be a more sensitive parameter than pore size. Starch as a minor component increases the median pore diameter in the sodium chloride matrix and consequently decreases the tablet strength. Furthermore, anisomorphous pore shapes lead to a severe reduction in strength. However, reduction of the particle size of the main component results in more spherical pores, making the effects of addition of a minor component less pronounced.

80

Effect of particle size on matrix strength

Chapter 5

5.4
1 2 3 4 5 6

References
Rumpf, H., Zur theorie der zugfestigkeit von agglomeraten bei krafbertragung an kontaktpunkten. Chemie-Ing.-Techn., 42 (1970) 538-540. Inglis, C.E., Stresses in a plate due to the presence of cracks and sharp corners. Roy. Inst. Naval Architects Trans., 55 (1913) 219-230. Griffith, A.A., Phenomena of rupture and flow in solids. Phil. Trans. Roy. Soc., A221 (1920) 163-168. Griffith, A.A., Theory of rupture. Proc. 1st Int. Conf. Appl. Mech., J. Waltman Jr., Delft (1924) 55. Hiestand, E.N., Principles, tenets and notions of tablet bonding and measurements of strength. Eur. J. Pharm. Biopharm., 44 (1997) 229-242. Caraballo, I., Fernndez-Arvalo, M., Holgado, M.A., Rabasco, A.M., Percolation theory: application to the study of the release behaviour from inert matrix systems. Int. J. Pharm., 96 (1993) 175-181. Fell, J.T., Newton, J.M., The tensile strength of lactose tablets. J. Pharm. Pharmacol., 20 (1968) 657-658. Duckworth, W.H., Discussion of Ryshkewitch paper by Winston Duckworth. J. Am. Ceram. Soc., 36 (1953) 68. Ryshkewitch, E., Compression strength of porous sintered alumina and zirconia. J. Am. Ceram. Soc., 36 (1953) 65-68. Eriksson, M., Alderborn, G., Mechanisms for post-compaction changes in tensile strength of sodium chloride compacts prepared from particles of different dimensions. Int. J. Pharm., 109 (1994) 59-72. Juppo, A.M., Relationship between breaking force and pore structure of lactose, glucose and mannitol tablets. Int. J. Pharm., 127 (1996) 95-102. Liu, D-M., Influence of porosity and pore size on the compressive strength of porous hydroxyapatite ceramic. Ceram. Int., 23 (1997) 135-139.

7 8 9 10

11 12

81

Chapter 5

Effect of particle size on matrix strength

82

A practical approach to predict mechanical properties

Chapter 6

Chapter 6 A practical approach towards a fit to predict mechanical properties of compacts containing two components

Summary The mechanical properties of tablets compacted from blends are not directly predictable from the mechanical properties of the individual components. This chapter introduces a structure dependent fit which gives a prediction of two major mechanical properties (elastic modulus and tensile strength) of compacts consisting of two particulate components. The structure dependent fit takes into account the properties of the individual materials, macroscopic changes in the compact structure and, if necessary, an interfacial interaction between particles of both components. Firstly, the relationship between the solid volume fraction and mechanical properties of compacts of the single materials is fitted by the percolation theory. From these fits the characteristics of the individual materials, i.e. the mechanical properties at zero porosity and the change of the volume fraction of particles that are coupled parallel in the direction of the acting force, are obtained. The material characteristics are led into the structure dependent fit to predict the maximal values (the upper boundary) and minimal values (the lower boundary) of the mechanical property of compacts made from the blends. It will be shown that the elastic modulus of the blends can be predicted by the upper boundary of the structure dependent fit. The calculated tensile strengths result into a relative estimation of the adhesive interfacial interaction between particles of both components. In the end, all steps of the structure dependent fit are included into a block diagram for the prediction of mechanical properties of compacts consisting of a blend. Keywords: compact; blend; structure dependent fit; percolation; adhesive interfacial interaction.

Submitted for publication in Powder Techn.

83

Chapter 6

A practical approach to predict mechanical properties

6.1

Introduction

Compacts produced in the pharmaceutical and food industry commonly consist of more than one component. Whereas data on single materials are generally available, numerous (time-consuming) experiments need to be performed to find good formulations for tablets compacted from blends. Therefore, a model or even a suitable fit to predict the mechanical properties of tablets containing more than one component is very valuable. Different authors, e.g. Thorton and Ning (1), Fortsyth and Rhodes (2), already have given powerful microscopic descriptions of particle-particle interactions influencing the mechanical properties of particulate systems. These models are based on fundamental descriptions giving an in-depth understanding, but the parameters in these models are often difficult to quantify. Additionally, the number of particles taken into consideration is low. Therefore, a more macroscopic method is proposed to predict the mechanical properties of blends directly from compacts compressed from single materials. The structure dependent fit introduced in this article is developed for tablets compressed from binary mixtures. It is based on a model of Kolarik (3-7) that describes the mechanical properties of polymer blends. Most models in polymer science only represent the properties of particulate composites. However, these models are not able to predict the mechanical properties of binary blends in which both components form a continuous system (matrix) inside the material structure. Kolariks description takes this bi-continuous structure into account. Since several articles already show that bi-continuous structures also exist in compacts compressed from binary powder mixtures (e.g. 8, 9, Ch. 4, Ch. 5), the model of Kolarik was taken as a basis for a structure dependent fit describing the different mechanical properties of tablets compressed from binary mixtures. To be useful in the prediction of mechanical properties of tablets compressed from binary mixtures, the structure dependent fit has to allow for: The properties of the single materials. An interval of mixtures in which both components form a continuous structure in the compact. Interfacial interaction between particles of both components (if applicable).

Mechanical properties of compacts are measured by different techniques. Many of these techniques measure the change in mechanical property as a function of the compacts porosity. Extrapolation of these values to zero porosity is thought to yield the mechanical property of a material, thereby avoiding the effects of pores and

84

A practical approach to predict mechanical properties

Chapter 6

internal changes on the compact structure. The mechanical property of a binary mixture at zero air fraction must instead be seen as the result of the composite formed by particles of both materials (10). The objective of this study is to introduce a suitable fit by which mechanical properties (elastic modulus and tensile strength) of compacts compressed from binary mixtures can be predicted from the mechanical properties of the individual components. Sodium chloride and pregelatinised starch were used as model components.

6.2 6.2.1

Model description Structure dependent fit

The structure dependent fit proposed in this article is built up from the two possible concepts of the particle structure of a material consisting of two components. The two components can be coupled in parallel (Fig. 6.1a) or in series (Fig. 6.1b). The parallel combination of the two components is also called the isostrain model because it assumes that the strains in the parallel components are equal. When the components are coupled in series (Fig. 6.1b), the components are discontinuous in the direction of the force. This is called the isostress model because of the equal stresses in the compounds.

1 1 2 2
V1p V2p

V1s V2s

Fig. 6.1. Different concepts of the particle structure of materials consisting of two components: (A) Parallel coupling of two components in a material structure. (B) Series coupling of two components in a material structure. (C) Schematic representation of the structure dependent fit. V is the volume fraction for each component, 1 or 2, coupled in parallel (subscript p) or in series (subscript s).

85

Chapter 6

A practical approach to predict mechanical properties

However, neither isostrain nor isostress models take into account bi-continuous structures or structures of a continuous component (matrix) combined with a discontinuous second component. To represent more complex behaviour of mechanical properties of binary mixtures, these two models are combined together in an equivalent box model (3) that represents a structure dependent calculation of the mechanical properties of binary polymer blends. A representation of this structure dependent fit is drawn in Fig. 6.1c. The blocks represent, for each component, which volume fraction (V) is coupled either parallel (subscript p) or in series (subscript s) in the direction of the active force. Each block is presumed to have the mechanical property of the component it represents in relation to its volume fraction. The parallel and in series branches are coupled parallel, so that the schematic representation will be in accordance with the actual compact structure. According to the structure dependent fit, the value of mechanical property Y of the blend (subscript b) is the result of the parallel and the series branch:

Yb = Y pV p + YsVs

(1)

The property Y of the parallel and the series branches is expressed by the following equations:

Yp = Ys =

Y1V1 p + Y2V2 p Vp Vs V1s V2 s + Y1 Y2

(2a)

(2b)

Combining equations 1, 2a and 2b, the property Y of the blend can be written as:

Yb = (Y1V1 p + Y2V2 p ) +

Vs V1s V2 s + Y1 Y2

Vs

(3)

with V1p+V2p+V1s+V2s = Vp+Vs = V1+V2 = 1. 6.2.2 Limits of property Y

In the original design of the structure dependent fit, it is expected that a complete transmission of stress is achieved between the particles of the different components. When the interfacial interaction is strong enough to transmit the stress between the
86

A practical approach to predict mechanical properties

Chapter 6

two different components, both the parallel and the series branch influence property Y. In this case, equation 3 is valid and can be regarded as an upper boundary Yb+. When the value of property Y of the first component is much higher than that of the second component and these components are coupled in series combined with complete stress transmission, the series branch will have the mechanical property Y2 (or Y1 in a reversed situation, whichever is the lowest). If Y2 is much smaller than Y1, equation 3 will transform to:

Yb + = (Y1V1 p + Y2V2 p ) + Y2Vs

(4)

It is, however, possible that the interfacial interaction between particles of the two different components is so low that any transmission of stress is impossible. As a consequence, the series branch will not contribute to the value of property Y. The mechanical property is only determined by the parallel branch and can be regarded as the lower boundary of property Y. The lower boundary Yb- is given by:

Yb = Y1V1 p + Y2V2 p

(5)

This phenomenon of low interfacial interaction and even particle disattachment during measurement is defined as dewetting or debonding in polymer science (5,11). Since equations 4 and 5 have a similar structure, the interaction factor A can be implemented in a combination of these two equations:

Yb = (Y1V1 p + Y2V2 p ) + AY2Vs

(6)

in which A = 0 stands for absence of interfacial interaction and A = 1 for a complete interfacial interaction. A value of A between 0 and 1 indicates an intermediate value of interfacial interaction between particles of the two components.

6.3

Materials and Methods

The model excipients used were sodium chloride (glidant-free, Akzo Nobel, Hengelo, The Netherlands) and pregelatinised potato starch (Paselli WA4, Avebe, Foxhol, The Netherlands). The 106-212 m sieve fraction of each material was obtained by sieving the powder with an Alpine Air Jet Sieve (Alpine, Augsburg, Germany). Sieve fractions were conditioned at 20C and 60% relative humidity for at least 4 days before mixing. Powder mixtures were prepared on a volume basis and mixed for 15 min at 90 rpm with a Turbula mixer (W.A. Bachofen, Basle, Switzerland).

87

Chapter 6

A practical approach to predict mechanical properties

Relative tapped densities of the single materials and the binary mixtures were calculated from the Heckel-plots (12) of compaction experiments with an average compaction speed of 3 mms-1. The relations between porosity and the tensile strength of tablets compressed from the single materials and each different mixture are the results from previous chapters. Relations between the porosity and the elastic modulus of compacts compressed from each mixture and single material were obtained by Dynamic Mechanical Analysis. Strips of 60 6 (1-2) mm, as described by Van der Voort Maarschalk et al. (13), were vibrated according to the three-point bending test. The maximum strain was 0.1% and the frequency was constant at 0.69 rads-1, which corresponds with an average compaction speed of 3 mms-1. For each different blend, at least eight strips with different porosities were measured.

6.4 6.4.1

Results and Discussion Tablets compressed from a single material

The structure of a tablet is often described by its volume fraction of the solid material. Particles of a component begin to establish a compact with a matrix of contacts at a certain volume fraction of that component, defined as the critical volume fraction or percolation threshold (Vcr). The network of contacts, i.e. the probability of bonding, is important for the coherence of the compact and the transmission of stress between the particles. The probability of bonding P can be reflected by the normalised solid volume fraction (14):

P=

V Vcr 1 Vcr

(7)

Property Y of a particulate structure is mainly the result of the probability of bonding between the particles. The change of property Y can be described by the percolation theory (15,16).

Y Y0 PT

(8)

where Y0 is the value of property Y at zero porosity and T is the critical exponent describing the change of property Y towards the percolation threshold.

88

A practical approach to predict mechanical properties

Chapter 6

Property Y is consequently related to the normalised volume fraction:

V Vcr Y = Y0 1V cr

(9)

Different values of T between 1.14 and 4.0 have been reported for conductivity (17,18), drug diffusion (19) and elastic modulus (20) of different particulate systems. In pharmaceutical research a value of T = 1 is often used in combination with the normalised solid volume fraction. The theoretical value of T of a three-dimensional structure consisting of round spheres is 2.0 (16). The anisomorphous shape of the particles and anisotropy of the compacts cause the wide deviation of the measured values of T around the theoretical value. In this study, the value of T has been fitted for the elastic modulus and tensile strength of the compacts compressed from the single materials. The elastic modulus and the tensile strength will not only be determined by the material, but also by air. Voids in the compacts are regarded as particles with an elastic modulus and tensile strength of zero. A tablet compressed from a single material can thus be considered as a composite in which the second component is air. Air in the shape of pores can form a percolating matrix through compacts or exist as finite clusters. Different authors (21, 22) have measured a volume fraction of around 0.10 for the percolation threshold of pores in tablets. The percolation thresholds of the single materials sodium chloride and pregelatinised starch were obtained from the densification profiles of the powders. The intercept with the -ln(porosity) axis of the Heckel-plots is regarded as the tapped density of the individual materials (14). Particles only form bonds when their positions are fixed at the tapped density, and increasing stresses cause particles to deform and enhance bond formation. The relative tapped densities of sodium chloride and pregelatinised starch were found to be 0.58 and 0.39, respectively, and were used as Vcr-values in equation 9.

89

Chapter 6

A practical approach to predict mechanical properties

16 Elastic modulus (GPa) 12 8 4 0 0.95

0.9

0.85

0.8

0.75

0.7

0.65

Solid volume fraction (-)

Fig. 6.2a. Measured elastic modulus versus solid volume fraction of () sodium chloride and () pregelatinised starch, respectively. Dotted lines represent the best fits obtained by equation 9.
9 8 7 6 5 4 3 2 1 0 1 0.9 0.8 0.7 0.6 0.5 Solid volume fraction (-)

Fig. 6.2b. Measured tensile strength versus solid volume fraction of () sodium chloride and () pregelatinised starch, respectively. Dotted lines represent the best fits obtained by equation 9.

Fig. 6.2a and Fig. 6.2b depict the best fit between the solid volume fractions and the elastic modulus and tensile strength for both sodium chloride and pregelatinised starch using equation 9. The maximum elastic modulus (E0), maximum tensile strength (S0) and values of T of both materials were obtained from these fits and are listed in Table 6.1. Sodium chloride is a rigid material with a high elastic modulus, whereas the softer starch particles have a much lower elastic modulus. According to

Tensile strength (MPa)

90

A practical approach to predict mechanical properties

Chapter 6

Table 6.1, the extrapolated tensile strengths of both materials are rather similar. The values of T for the solid volume fraction-tensile strength relations are higher than those obtained for the elastic modulus. This illustrates the higher influence of the compact structure on the change in tablet strength.
Table 6.1. Fitted values of the elastic modulus (6.1a) and tensile strength (6.1b) of compacts compressed from the single materials sodium chloride and pregelatinised starch by equation 9. Table 6.1a Elastic modulus Sodium chloride Pregelatinised starch Table 6.1b Tensile strength Sodium chloride Pregelatinised starch

E0 (GPa) 18.4 2.68

T 1.12 2.11

r2 0.990 0.961

S0 (MPa) 11.5 12.3

T 2.45 2.92

r2 0.971 0.996

In a fully dense compact (zero air fraction) compressed from a single material, all particles are coupled to each other and determine the maximum value of property Y. When the compact contains air, it consists of two components. It becomes possible to model the structure of the particles and the air as described in sec. 6.2.1. These two components can be coupled in parallel, in series and as a combination of these two branches (Fig. 6.1c). Note that air gives no direct contribution to either the elastic modulus or tensile strength. The implication is that a compact can only have a nonzero value of modulus or strength when the particles of the solid component are coupled in parallel, while the in series branch is always zero. This implies that the value of Y is entirely determined by the parallel contribution of the material and equation 3 collapses to Y = Y0Vp. From this perspective, the parallel volume fractions of the material can be calculated from:

V Vcr Vp = 1V cr

(10)

After the calculations of the parallel volume fractions by equation 10, the other volume fractions, e.g. Vs, can be subsequently calculated (see subscript eq. 3) in order to be used in calculations of the elastic modulus or tensile strength of compacts consisting of two components. Fitting the solid volume fraction-mechanical property relation with equation 9 enables three different variables (i.e. Y0, T and Vp) to be calculated for each single
91

Chapter 6

A practical approach to predict mechanical properties

material. These variables and the relative tapped density are easy to obtain from single material measurements. The approach is to implement these input data into the calculations used for the predictions of the mechanical properties of compacts consisting of two component powder blends. The outcomes of the structure dependent fit (eq. 3-5) are predicted, i.e. not fitted, values using the individual values obtained from the single materials. 6.4.2 Elastic modulus of binary mixtures

During compaction of blends of sodium chloride and pregelatinised starch, particles of both materials will deform plastically (23). The ability to form a matrix in the compact depends on the difference in yield pressure (Py) between the two materials. As observed in Ch. 2 and Ch. 3, the starch particles deform greatly, whereas the sodium chloride particles hardly change their shape during densification of the mixtures. For this reason, the starch particles are considered to act as air during densification of a mixture with sodium chloride particles and to form a matrix already at low solid volume fractions. Hence, the earlier mentioned percolation threshold of 0.10 of the air (porosity) (21,22) instead of the original value of 0.39 (sec. 6.4.1) has been used in the structure dependent fit as the percolation threshold of starch. For sodium chloride the original percolation threshold of 0.58 has been maintained. The elastic modulus, tensile strength and T value of the single component data (Table 6.1) were applied in the calculations of eq. 3-5 to predict the values of the blends. The elastic modulus of compacts is measured at very low strains (< 1%) to remain in the linear region of the strain-stress relationship. The applied strain in this study is so low that multicomponent particulate systems are normally characterised by complete transfer of stress through the particles of the different materials (24). From this point of view, the upper boundary equations (eq. 3 and eq. 4) of the structure dependent fit represent the elastic moduli of binary mixtures. The values of E0 of the mixtures calculated by the structure dependent fit are given in Fig. 6.3a. The dashed and continuous lines are the upper boundaries given by eqs. 3 and 4, respectively, and give the highest predicted values of the E0 of the blends. The lower boundary (dotted line, eq. 5) displays a minimal value at a sodium chloride volume fraction of 0.60. These predicted upper and lower boundaries are subsequently compared with measured E0 of the blends.

92

A practical approach to predict mechanical properties

Chapter 6

20 Elastic modulus (GPa) 15 10 5 0 1 0.8 0.6 0.4 0.2 0 Sodium chloride volume fraction (-)

Fig. 6.3a. Predicted elastic modulus of compacts compressed from blends with zero porosity. Dashed, continuous and dotted lines represent the outcome of equations (3), (4) and (5), respectively. (Input data: VNaCl,cr = 0.58, VStarch,cr = 0.10, TNaCl = 1.12 and TStarch = 2.11).

For every different blend consisting of sodium chloride and pregelatinised starch, the solid volume fraction-elastic modulus relation was first fitted by equation 9 to obtain the maximum elastic modulus at zero porosity. The required relative tapped densities necessary were measured for each different blend by the Heckel-plots and are given in table 6.2 together with the fit results.
Table 6.2. Fitted values of the elastic modulus versus solid volume fraction of compacts compressed from blends of sodium chloride and pregelatinised starch by equation 9. Sodium chloride volume Relative tapped fraction in mixture (-) density (-) 1.00 0.58 0.975 0.59 0.95 0.59 0.925 0.60 0.90 0.57 0.85 0.59 0.80 0.59 0.75 0.58 0.70 0.57 0.60 0.54 0.40 0.46 0.20 0.46 0.00 0.39 E0 (GPa) 18.4 18.3 16.9 15.9 15.6 16.3 15.9 11.5 10.0 6.68 3.13 2.45 2.68 T (-) 1.12 1.22 1.43 1.62 2.57 3.51 3.97 4.11 3.99 4.46 3.28 2.18 2.11 r2 (-) 0.990 0.925 0.948 0.800 0.824 0.886 0.676 0.890 0.833 0.929 0.922 0.934 0.961

93

Chapter 6

A practical approach to predict mechanical properties

The fact that beams consisting of binary mixtures show greater fluctuations in the solid volume fraction-elastic modulus relation than beams compressed from the single materials is reflected by the lower coefficients of correlations. The extrapolated E0 of the compacts of the blends are given in Fig. 6.3b together with the predictive boundaries of the structure dependent fit.
20 Elastic modulus (GPa) 15 10 5 0 1 0.8 0.6 0.4 0.2 0 Sodium chloride volume fraction (-)

Fig. 6.3b. Measured elastic modulus of () compacts compressed from blends with zero porosity. Dashed, continuous and dotted lines represent the outcome of equations (3), (4) and (5), respectively (similar as in Fig. 6.3a.).

The calculated lower boundary (eq. 5) shows the largest deviation from the extrapolated E0 of the mixtures. The calculated upper boundaries (eq. 3 and eq. 4) correspond better with the measured data. The calculated upper boundary equation 3 appears to give the best fit of the extrapolated E0 of the blends. This proves that (1) the interfacial interaction is strong enough to transmit all stress between particles of different materials during elastic modulus measurements and that (2) particles of both materials influence the contribution of the in series branch of the structure dependent fit (Fig. 6.1c). Therefore, the upper boundary equation 3 of the structure dependent fit can be used to predict the elastic modulus of binary mixtures at zero porosity from the elastic moduli of the individual materials.

94

A practical approach to predict mechanical properties

Chapter 6

6.4.3

Tensile strength of binary mixtures

According to fracture mechanics, the fracture type of a tablet is independent of the material it is composed of. Tablets fracture in a brittle fashion (25). The tensile strength of compacts is affected by pores or cracks present inside the compact. Pores partially control the compact strength as an effect of their ability to concentrate stress at the tips (26, 27). However, by extrapolating the tensile strengths of compacts compressed from binary mixtures to zero porosity, stress concentrations as a result of the presence of pores are considered not to be present inside the structure. In contrast to measurements of the elastic modulus, tensile strength measurements are conducted at higher strains. In this situation, it is important whether the interfacial interaction between particles of the different materials is strong enough to transmit all acting stress. The interfacial interaction determines the contribution of the series branch of the structure dependent fit to the total tensile strength of the blend. The predicted upper boundaries (eq. 3 and eq. 4) and lower boundary (eq. 5) of the tensile strength of the blends are calculated using the tensile strength data given in Table 6.1 and are shown in Fig. 6.4. The upper boundaries equations 3 and 4 predict comparable and high tensile strengths of the mixtures, whereas the lower boundary equation 5 passes through a minimum at a sodium chloride fraction of about 0.60.
16 Tensile strength (MPa) 12 8 4 0 1 0.8 0.6 0.4 0.2 0 Sodium chloride volume fraction (-)

Fig. 6.4. Measured tensile strengths of () compacts compressed from binary mixtures of sodium chloride and pregelatinised starch with zero porosity. Dashed, continuous and dotted lines are the predictive results of equations (3), (4) and (5), respectively. (Input data: VNaCl,cr = 0.58, VStarch,cr = 0.10, TNaCl = 2.45 and TStarch = 2.92).

95

Chapter 6

A practical approach to predict mechanical properties

Similar to the elastic modulus calculations, the extrapolated tensile strengths at zero porosity of compacts of the blends were determined by fitting the solid volume fraction-tensile strength relationships using equation 9. The results are given in Table 6.3. Fig. 6.4 depicts these extrapolated tensile strengths at zero porosity of tablets compressed from binary mixtures of sodium chloride and starch. The tensile strength of the tablets passes a minimum as a function of the mixture composition. It is clear that the upper boundaries in Fig. 6.4 (eq. 3 and 4), which are mathematically not capable of passing through a minimum, are unsuitable to predict the change in tensile strength of the mixtures. The lower boundary (eq. 5), on the other hand, has a minimum in the tensile strength-composition relationship and its predicted values are lower than the tensile strength of the tablets compressed from the mixtures.
Table 6.3. Fitted values of the tensile strength versus solid volume fraction of compacts compressed from blends of sodium chloride and pregelatinised starch by equation 9. Sodium chloride volume fraction in mixture (-) 1.00 0.975 0.95 0.925 0.90 0.85 0.80 0.75 0.70 0.60 0.40 0.20 0.00 S0 (MPa) 11.50 11.05 11.72 11.34 9.55 8.51 7.95 6.04 6.46 5.38 5.85 9.41 12.30 T (-) 2.45 4.01 4.97 5.45 5.62 4.40 4.40 3.74 3.87 3.28 2.80 2.61 2.92 r2 (-) 0.971 0.982 0.979 0.961 0.952 0.959 0.923 0.961 0.943 0.942 0.969 0.986 0.996

As the experimentally determined tensile strengths of the blends are between the upper boundaries and the lower boundary, it can be concluded that no complete stress transmission has taken place. No complete stress transmission indicates that the interaction factor A of equation 6 has a value between 0 and 1. So, the conclusion is that adhesion plays a significant role. To determine the value of A, an additional measurement of one particular blend has to be conducted. When the contribution of the series branch is high, i.e. AY2Vs is (relatively) large, an accurate estimation of the A-value can be made. The tensile strength of the mixture composition with the lowest tensile strength is almost completely determined by the series contribution. This

96

A practical approach to predict mechanical properties

Chapter 6

mixture composition should be calculated using the derivative of equation 6. As the value of A is not known, it is impossible to calculate the composition that gives the weakest compacts. Instead, the lowest contribution of the parallel branch to the total tensile strength is given by the derivative of the lower boundary (eq. 5). This minimum of the lower boundary is expected to result into a similar mixture composition as that given by equation 6. Equation 5 has its minimum when:

V1 V1cr 1 V 1cr

T1 1

1 V1 V2 cr 1V 2 cr

T2 1

Y2T2 1 V1cr Y1T1 1 V2 cr

(11)

Using the input data of Table 6.1 and the critical volume fractions of both components into equation 11, it was calculated that the mixture composition with the lowest contribution of the parallel branch has a sodium chloride volume fraction (V1) of approximately 0.6. Consequently, the A factor is calculated by subtracting the parallel branch (eq. 5) from the extrapolated tensile strength at zero porosity of this mixture composition. This difference equals AY2Vs in which Y2 is the tensile strength of the weakest component. Fig. 6.5a shows tensile strength-compact composition relationships using different values of A.
16 Tensile strength (MPa) 12

0.80
8 4 0 1 0.8 0.6 0.4 0.2 0 Sodium chloride volume fraction (-)

0.45 0.10

Fig. 6.5a. Measured tensile strength of () tablets compressed from sodium chloride, pregelatinised starch and a blend with a sodium chloride volume fraction of 0.60 (all with zero porosity). Continuous lines show the contributions of the in series branch in equation 6 with different values of A to the total tablet strength.

97

Chapter 6

A practical approach to predict mechanical properties

The value of the interaction factor A is thought to reflect the adhesive interfacial interaction between particles of sodium chloride and pregelatinised starch (sec. 6.2.2). The A-value of 0.45 has been calculated as the relative contribution of the in series branch of the mixture containing a sodium chloride volume fraction of 0.60. Therefore, the adhesive interfacial interaction is thought to be about half as strong than the weakest cohesive bonding which is the sodium chloride bonding in these blends. Finally, Fig. 6.5b shows the predicted strengths of compacts from mixtures using an A-value of 0.45 as described by equation 6. All measured tensile strengths of the different mixture compositions show a good correlation with the calculated values. So, the tensile strength of the mixtures can be predicted from the individual tensile strengths with one additional measurement of the tensile strength of the blend with the lowest contribution of the parallel branch.
16 Tensile strength (MPa) 12 8 4 0 1 0.8 0.6 0.4 0.2 0 Sodium chloride volume fraction (-)

Fig. 6.5b. Measured tensile strength of () tablets compressed from binary mixtures of sodium chloride and pregelatinised starch with zero porosity in comparison with equation 6 with A = 0.45 (continuous line) ( used in fit).

6.4.4

Block diagram

The structure dependent fit presented in this paper promises to be useful tool to predict mechanical properties of compacts compressed from binary mixtures. To make the different steps clearer, all steps are included into a block diagram (Fig. 6.6). It is not yet possible to use such a block diagram for all types of blends. This is because there is the assumption that the yield pressures of the materials are different. After collecting all parameters of the single materials (Y0, Vcr, T, Py) as mentioned at the start of the block diagram, the difference in yield pressure determines the matrix formation

98

A practical approach to predict mechanical properties

Chapter 6

of the different components in the compact. The strains applied during the different types of measurements are important for the choice of the appropriate equation of the structure dependent fit. Equation 3 seems to be the most suitable equation for the elastic modulus of blends and equation 6 for the tensile strength of blends. In formulation practice this approach can help to select components that enable coherent compacts to be made. Additionally, the method allows quantifying the adhesion between solid materials.
Collect Y0, V cr, T, Py of single materials

Py2<<Py1
(and equal particle sizes)

No prediction possible, yet

+
V1cr = V 1cr V2cr = 0.10

Complete stress transfer

Mixture value by equation 3

Calculation of lowest value equation 11

Additional mixture measurement

Calculation adhesive interaction A

Mixture value by equation 6

Fig. 6.6. Block diagram to select appropriate equation for mixture value calculation of mechanical property by the structure dependent fit.

99

Chapter 6

A practical approach to predict mechanical properties

6.5 A E Py S T V Y

Nomenclature Interfacial interaction Elastic modulus Yield pressure Tensile strength Critical universal exponent Volume fraction Mechanical property

Subscripts: b cr p s 0 1 2 + 6.6
1 2

blend critical parallel in series zero porosity component 1 component 2 upper boundary lower boundary References
Thorton, C., Ning, Z., A theoretical model for the stick/bounce behaviour of adhesive, elastic-plastic spheres. Powder Techn., 99 (1998) 154-162. Forsyth, A.J., Rhodes, M.J., A simple model incorporating the effects of deformation and asperities into the Van der Waals force for macroscopic spherical solid particles. J. Colloid Interface Sci., 223(1) (2000) 133-138. Kolarik, J., A model for the yield strength of binary blends of thermoplastics. Polymer, (17) (1994) 3631-3637. Kolarik, J., Prediction of the yield strength of polymer blends. Polym. Networks Blends, 5(2) (1995) 87-93. Kolarik, J., Evaluation of the extent of interfacial debonding in polymer blends. Polymer, 37(6) (1996) 887-891. Kolarik, J., Three-dimensional models for predicting the modulus and yield strength of polymer blends, foams, and particulate composites. Polym. Compos., 18(4) (1997) 433441. Kolarik, J., Simultaneous prediction of the modulus, tensile strength and gas permeability of binary polymer blends. Eur. Polym. J., 34(5/6) (1998) 585-590. Blattner, D., Kolb, M., Leuenberger H., Percolation theory and compactibility of binary powder systems. Pharm. Res., 7(2) (1990) 113-117.

3 4 5 6

7 8

100

A practical approach to predict mechanical properties 9

Chapter 6

Holman, L.E., Leuenberger, H., The effect of varying the composition of binary powder mixtures and compacts on their properties: A percolation phenomenon. Powder Techn., 60 (1990) 249-258. Bassam, F., York, P., Rowe, R.C., Roberts, R.J., The Youngs modulus of binary powder mixtures. Powder Techn., 65 (1991) 103-111. Errico, M., Liquid coating processes. In: R.J. Stokes, D.F. Evans, Fundamentals of interfacial engineering, Wiley-VCH, New York, NY, USA, 1997. Heckel, R.W., Density-pressure relationships in powder compaction. Trans. Metall. Soc. AIME., 221 (1961) 671-675. Van der Voort Maarschalk, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., The effect of viscoelasticity and tabletting speed on consolidation and relaxation of a viscoelastic material. Eur. J. Pharm. Biopharm., 42 (1996) 49-55. Holman, L.E., Leuenberger, H., The significance of slopes of the semilogarithmic relationship between hardness and solid fraction of porous compacts. Powder Techn., 64 (1991) 233-247. De Gennes, P.G., On a relation between percolation theory and the elasticity of gels. J. Phys. Lett. (Paris), 37 (1976) L1. Stauffer, D., Introduction to percolation theory, Taylor & Francis, London (1985) 3-103. Roldughin, V.I., Vysotskii, V.V., Percolation properties of metal-filled polymer films, structure and mechanisms of conductivity. Prog. Org. Coatings, 39 (2000) 81-100. Thommerel, E., Valmette, J.C., Musso, J., Villain, S., Gavarri, J.R., Spada, D., Relations between microstructure, electrical percolation and corrosion in metal-insulator composites. Mater. Sci. Eng., A328 (2002) 67-79. Bonny, J.D., Leuenberger, H., Matrix type controlled release systems: I. Effect of percolation on drug dissolution kinetics. Pharm. Acta Helv., 66(5/6) (1991) 160-164. Kuentz, M., Leuenberger, H., Kolb, M., Elasticity of polymer tablets considered as a network of contacts. S.T.P. Pharma Sci., 9(6) (1999) 531-538. Holman, L.E., The compaction behaviour of particulate materials. An elucidation based on percolation theory. Powder Techn., 66 (1991) 265-280. Steendam, R., Eissens, A.C., Frijlink, H.W., Lerk, C.F., Plasticisation of amylodextrin by moisture: Consequences for drug release from tablets. Int. J. Pharm., 204 (2000) 23-33. Van der Voort Maarschalk, K., Zuurman, K., Vromans, H., Bolhuis, G.K., Lerk, C.F., Stress relaxation of compacts produced from viscoelastic materials. Int. J. Pharm., 151 (1997) 27-34. Hill, D., An introduction to composite materials, Cambridge University Press, Cambridge, England (1981). Stanley, P., Mechanical strength testing of compacted powders. Int. J. Pharm., 227 (2001) 27-38. Griffith, A.A., Phenomena of rupture and flow in solids. Phil. Trans. Roy. Soc., A221 (1920) 163-168. Griffith, A.A., Theory of rupture. Proc. 1st Int. Conf. Appl. Mech., J. Waltman Jr., Delft (1924) 55.

10 11 12 13

14

15 16 17 18

19 20 21 22 23

24 25 26 27

101

Chapter 6

A practical approach to predict mechanical properties

102

Conclusions and future research

Chapter 7

Chapter 7 Conclusions and future research

7.1

Conclusions

Chapters 2-5 of this thesis describe the compaction cycle of powder blends consisting of two model materials. In the chapters, the interactions between the materials that occur during compaction are related to the changes in tablet strength. To understand the tablet strength as a result of interactions between multiple components, several aspects of compact structure and strength have been investigated. Both model materials (sodium chloride and pregelatinised starch) have ductile deformation characteristics, so the densification of the powder bed is mainly achieved through particle deformation. Among the many differences in mechanical and physical properties between the materials, the difference in yield pressure is of paramount importance. The binary powder blends can be regarded as mixtures of a soft and a rigid material. When different particle ratios of a soft and a rigid ductile material are compressed, a linear relationship is obtained between the measured yield pressure and the mixture composition. However, small negative deviations from this linearity were discovered in this study. When a rigid particle is completely surrounded by softer particles during compression, the latter will to a large extent protect this rigid particle from being affected by the applied load. The relaxation behaviour of binary powder blends differs from that of the individual components. Expressing the tablet relaxation as a porosity expansion showed that the mixtures had higher tablet relaxations than was expected from linear interpolation of the relaxation behaviours of the individual materials. This phenomenon was defined as the extra porosity expansion. Since tablet relaxation is counteracted by interparticle bonding, this behaviour is attributed to the fact that adhesion forces are lower than cohesion forces between the particles in the blend. Interparticle bonding influences the tablet strength: lower adhesion forces reduce the tablet strength. In fact, the extra porosity expansion is the first indication of a decrease in tablet strength of tablets compressed from mixtures compared to those prepared from single materials. The link between the compaction behaviour of binary powder blends and the change in tablet strength is to be found in the tablet structure. In this study, porosity
103

Chapter 7

Conclusions and future research

has been used as an independent variable to describe tablet structure and strength. Burning out of the softer particles from the tablets allowed the assessment of the tablet structure formed by the rigid material alone. The difference in pore size distribution before and after heat treatment implies that the softer particles are greatly deformed between the rigid particles due to the difference in yield pressure. This finding is further supported by Scanning Electron Microscopy-observations. The removal of the softer particles made it possible to determine the percolation threshold of the rigid material. This threshold is found at relatively high concentrations and is related to its particle size. The percolation threshold of the soft material remains similarly at low concentrations and is hardly influenced by the particle size of the rigid material. The establishment of a matrix already at low concentrations by the soft component is explained by the large particle deformation of this component. The following conclusions could be drawn on the strength of the matrix of rigid particles after removal of the softer particles from the tablets compressed from the blends: The strength of the matrix only is comparable to the original tablet strength when the particles of the soft component are present in the tablet structure as clusters. Hence, it can be concluded that the matrix strength of the rigid material determines the tablet strength in this composition range. The strength of the rigid material matrix is lower than the tablet strength of the mixtures when two continuous networks coexist, i.e. at concentrations between both percolation thresholds.

Besides a particulate structure, tablets also contain pores. In tablets compressed from binary mixtures used in this study, alterations caused by variations in the tablet composition are observed in the larger pore range of the pore size distribution while the smaller pores stay constant in number and size. The median pore diameter appears to be a suitable parameter to reflect the changes in larger pores. The larger pores in the structure of tablets compressed from blends are primarily the result of tablet relaxation. Pores in materials are known to act as stress concentrators during strength measurements. Generally, large pores induce a higher decrease in tablet strength than small pores. The addition of softer particles to the matrix consisting of the rigid particles results into an increase of the median pore diameter. The creation of larger pores in the matrix structure is the most important reason for the decrease in strength of tablets containing high amounts of the rigid material. It was found that the median pore diameter shows a unique relationship with the tablet strength. The strength of
104

Conclusions and future research

Chapter 7

tablets compressed only from particles of the rigid material is directly related to the median pore diameter. Tablets compressed from blends where the soft component does not give a continuous structure were also found to have a unique, but different, relationship between median pore diameter and tablet strength. Both relationships are independent of the particle size and the concentration of the rigid material. However, the presence of softer particles during compaction of the blends increases the negative effect of large pores on the tablet strength. The difference between both relationships is because also pore shape is important. The pore shape was defined as a length-todiameter ratio of ellipsoidal holes. The larger pores in tablets compressed from the blends are longer and have sharper tips than the more spherically shaped pores in the tablets compressed from only the rigid material. The pore shape in the tablets compressed from blends contributes to a larger reduction of the tablet strength. In Chapter 6, the structure dependent fit is introduced to predict mechanical properties of compacts consisting of binary mixtures from the individual material properties. The structure dependent fit also calculates a relative interfacial interaction between particles of both components from the prediction of the strength of compacts produced from the mixtures. The adhesion forces between sodium chloride and starch are estimated to be the half of the cohesion forces of sodium chloride.

7.2

Future research

The research and results presented in this thesis offer a new approach towards compaction studies of powder blends. Although this study only has investigated blends that consist of two selected ductile materials, the principles also apply to other blends consisting of ductile materials. Moreover, this research on changes in compact structure of blends provides not only a better understanding of the tablet strength, but can also be used to understand other physical properties of tablets, such as tablet disintegration, drug diffusion during dissolution tests and tablet hardness. Future research should be performed to gain more knowledge on important aspects of the relationship between the compaction behaviour of binary blends and the final tablet strength. Other types of blends that contain combinations of ductile/brittle and brittle/brittle materials should be investigated. These blends can also be characterised by the basic material properties of two components (e.g. yield/fracture strength, type of particle bonding, particle size distribution, etc.). Even more important are the effects of the adhesive particle interactions between different materials on tablet relaxation, pore formation and changes in tablet strength. The porosity as

105

Chapter 7

Conclusions and future research

variable has to be further investigated, even though this thesis shows (chapter 4 and 5) that the median pore diameter appears to be a universal parameter in relation to tablet strength. In this study, the decrease in tablet strength is only described for matrix systems of the rigid material with clusters of soft particles. The influences of particles of a rigid material on the matrix structure and strength of softer particles should be further investigated. The question, which interaction between the pore size and interparticle bonding determines the tablet strength at concentrations between both percolation thresholds, still remains unanswered. To solve this problem, it is necessary to gain more knowledge about fracture dynamics of tablets. To estimate the effect of pore size and pore shape on tablet strength, the perfect tablet strength of tablets without cracks and pores has to be calculated from the fracture properties of the material itself. Hence, more knowledge has to be obtained about the mechanism of diametrical compression failure of cylindrical objects. Other fields of research, such as mechanical engineering, polymer science and ceramics, could supply information for the solution of this problem. The structure dependent fit is the first step towards the prediction of the strength of a tablet compressed from a binary mixture (including porosity) from the individual material strengths. The structure dependent fit has to be tested for other binary mixtures and adjustments have to be made. When more knowledge is obtained about the relationship between adhesive particle interaction and the creation of larger pores in the tablet structure, it should be possible to insert the third component, air, into the model. In the future, the strength of tablets compressed from binary mixtures should not be a black box anymore, but an understandable and even predictable result of the compaction behaviour of powder blends.

106

Summary

Summary
The efficiency of a powder consisting of two different materials to be compacted into a solid dosage form, i.e. a tablet, does not only depend on the performance of the individual materials, but also on interactions between the materials during the compaction process of a powder blend. Chapter 1 presents a short overview of the compaction behaviour of single materials in general and of binary powder blends in particular. The densification behaviour of a single material can be classified as either brittle or ductile. The studies on binary powder blends that have been reported in the literature were divided into groups based on combinations of these individual densification behaviours. Because of the many variations in materials and experimental conditions, it is rather difficult to compare the results of different compaction studies. Therefore, the aim of this thesis was to establish relationships between the compaction behaviours of binary powder blends (containing two model materials) and the resulting tablet strength. Chapter 2 discusses the compaction behaviour of binary mixtures consisting of sodium chloride and pregelatinised starch. Both materials are known to deform plastically during densification. The yield pressures of the binary mixtures show an almost linear relationship with the mixture composition. As an effect of their lower yield pressure, starch particles yield earlier than sodium chloride particles during the densification process. The following enclosure of single sodium chloride particles by starch particles prevents these sodium chloride particles to yield or crack. The relaxation of the tablets compressed from blends is higher than the relaxation calculated by linear interpolation of the relaxation behaviours of the two pure materials. The difference between the measured porosity expansion and the data obtained by linear interpolation can be considered as a measure for the reduced interparticle bonding. SEM-photographs show large pores between the sodium chloride and pregelatinised starch particles. This observation indicates that the reduced interparticle bonding is caused by low adhesive forces. Since porosity is directly related to tablet strength, all tablets compressed from blends in this thesis are prepared with a constant porosity of 15%. The tensile strength of tablets compressed from the binary mixtures is not linearly related to the strength of tablets prepared from single materials; it shows a decreased tensile strength relative to interpolation. The measured decrease of the tensile strength of the tablets is also considered to be the result of reduced interparticle bonding. In chapter 2, it is shown that there exists a similar

107

Summary

relationship between the tensile strength reduction and the percentage of starch on the one hand and the extra porosity expansion and the starch percentage on the other hand. Chapter 3 describes the internal structure of tablets compressed from mixtures of sodium chloride and pregelatinised starch. The difference between the self-ignition temperature of starch at 300 C and the melting temperature of sodium chloride at 800 C enables the starch particles to be burned out, while the sodium chloride structure in the tablets is maintained. The minimum particle diameter of pregelatinised starch inside tablets compressed from mixtures was calculated from the difference between the initial pore size distribution and the pore size distribution after removal of the starch particles. Subsequently, the tablets were carefully crushed. The size distribution of these sodium chloride particles was measured by laser diffraction. It was found that the diameter of the sodium chloride particles had hardly changed, whereas the minimum diameter of starch particles strongly decreased during the compaction process. As an effect of the difference in yield pressure, the harder sodium chloride particles cause deformation of the softer starch particles, resulting in a change in particle shape. The pore size distribution of tablets compressed from mixtures of sodium chloride and starch is typically that of viscoelastic materials; the larger pores (> 5 m) change upon variations in the tablet composition, while the small pores stay constant in number and size. The median pore diameter in tablets compressed from the mixtures is higher than the median pore diameter in tablets compressed from the pure materials. This chapter shows that the formation of large pores is the result of the extra porosity expansion of tablets compressed from binary mixtures of sodium chloride and pregelatinised starch with equal particle sizes. Chapter 4 deals with the strengths of continuous networks, i.e. matrices, of sodium chloride particles in tablets compressed from binary mixtures of sodium chloride with low concentrations of pregelatinised starch and in pure sodium chloride tablets. In order to compare the different tablets, the tablet strength is reflected as a function of the sodium chloride volume fraction in the tablet. Starch particles in the tablets containing the blends decrease the sodium chloride volume fraction-tensile strength relationship compared with that of pure sodium chloride tablets. To determine the contribution of the sodium chloride matrix to the tablet strength, the starch particles were removed from the mixture tablets by heat treatment (described in Chapter 3). Determination of the strengths of these heat-treated tablets reveals that the sodium chloride matrix strength determines the tablet strength of mixture tablets that consist of a matrix of sodium chloride particles and clusters of starch particles. The decrease of the sodium chloride matrix density in the three different tablets (pure sodium chloride tablets, tablets containing the mixtures and heat-treated tablets)

108

Summary

coincides with an increase of the median pore size. The matrix in sodium chloride tablets shows a higher tensile strength to median pore size relation than the matrices in the mixture and heat-treated tablets. Based on calculations according to the theory of elastic-brittle fracture, it is suggested that the initial presence of starch particles during tablet compaction causes the pores in the matrices of the mixtures and heat-treated tablets to be relatively more flat and longer. These pores weaken the sodium chloride matrix in the mixture and heat-treated tablets to a larger extent, than the shorter, more spherical pores formed during compaction of pure sodium chloride. In Chapter 5, the material of main interest is sodium chloride consisting of different particle size distributions, while the other material is pregelatinised starch with a constant particle size distribution. Particle size and particle size distribution influence the mechanical strength of compacts. Tablets compressed from pure sodium chloride and from blends were prepared with a constant porosity of 15% and their tensile strengths were studied as a function of the sodium chloride volume fraction. The removal of the starch particles from the tablets compressed from blends made it possible to investigate the percolation thresholds of both materials and measure the strengths of the sodium chloride matrices consisting of different particle size fractions. The percolation threshold is the critical concentration of a material at which the continuous network turns into finite clusters. The percolation threshold of sodium chloride is directly related to its mean particle size, whereas the percolation threshold of starch is constant at low starch volume fractions. Furthermore, starch particles cause an increase in median pore diameter in tablets compressed from blends of which the sodium chloride matrix completely determines the tensile strength. Alterations in tensile strength of these mixture tablets are suggested to be related to the median pore diameter. A unique relationship is found between the median pore diameter and the tensile strength of tablets prepared from blends containing high amounts of sodium chloride. This relationship is independent of the sodium chloride mean particle size and sodium chloride volume fraction. A similar relationship exists for tablets compressed from only sodium chloride particles, but this relationship depicts higher tensile strengths. The deviation between both relations is explained by the difference in pore shape (Chapter 4). Chapters 2-5 describe the sequence of phenomena occuring during and after the compaction of binary powder blends consisting of sodium chloride and pregelatinised starch. The interactions between particles of both materials determine the final tablet structure. The chapters show that the creation of larger pores is the main cause of the reduction of the tablet strength. The median pore diameter appears to be a universal parameter to describe the strength of tablets containing different sodium chloride

109

Summary

matrices. Chapters 4 and 5 distinguish the effect of starch particles on the pores in sodium chloride matrices by the definition of the pore shape. The mechanical properties of tablets compacted from blends are generally not directly predictable from the mechanical properties of the individual components. Chapter 6 introduces the structure dependent fit which gives a prediction of two major mechanical properties (elastic modulus and tensile strength) of compacts consisting of two particulate components. The structure dependent fit takes into account the properties of the individual materials, macroscopic changes in the compact structure and, if necessary, an interfacial interaction between particles of both components. Firstly, the relationship between the solid volume fraction and mechanical properties of compacts of the single materials were fitted by the percolation theory. From these fits the characteristics of the individual materials, i.e. the mechanical properties at zero porosity and the change of the volume fraction of particles that are coupled parallel in the direction of the acting force, were obtained. The material characteristics were applied into the structure dependent fit to predict the maximal values (the upper boundary) and minimal values (the lower boundary) of the mechanical property of compacts made from the blends. It has been shown that the elastic modulus of the blends can be predicted by the upper boundary of the structure dependent fit. The calculated tensile strengths at zero porosity result into a relative estimation of the adhesive interfacial interaction between particles of both components. In the end, all steps of the structure dependent fit are included into a block diagram for the prediction of mechanical properties of compacts consisting of a blend. The final chapter of this thesis (Chapter 7) presents the conclusions based on the material properties of the two materials used in the studies. Furthermore, it presents suggestions for further research on the compaction behaviour of powder blends.

110

Samenvatting

Samenvatting
Het tablet is om vele redenen een populaire toedieningsvorm van geneesmiddelen. Het gebruikersgemak en het gemak waarmee ze grootschalig kunnen worden geproduceerd zijn slechts twee van de voornaamste redenen. Echter, een tablet is niet alleen een drager van een of meer geneesmiddelen, maar het moet ook voldoen aan vooropgestelde specificaties betreffende mechanische sterkte, doseernauwkeurigheid en de afgiftesnelheid van het geneesmiddel. Om aan deze eisen te voldoen is het noodzakelijk om in de formulering (de combinatie van alle materialen in het poedermengsel) functionele hulpstoffen toe te passen en het productieproces onder goed gecontroleerde omstandigheden uit te voeren. Aangezien verschillende materialen binden met verschillende bindingskrachten, komt het in de praktijk geregeld voor dat de uiteindelijke sterkte van de geslagen tabletten onvoldoende is. Het is tot op heden grotendeels onbekend op welke wijze de verschillende materialen elkaar betreffende binding benvloeden en de uiteindelijke tabletsterkte die daaruit resulteert. Een literatuuroverzicht over het onderzoek naar het tabletteergedrag van enkelvoudige materialen en poedermengsels wordt in hoofdstuk 1 weergegeven. Een goed uitgangspunt voor onderzoek naar interacties tussen verschillende materialen is de keuze van een poedermengsel dat slechts bestaat uit losse deeltjes van twee verschillende materialen. Het proces van tabletteren wordt ook wel compacteren genoemd; het maken van een compact. Het compactieproces kan verdeeld worden in twee opeenvolgende fasen: een verdichtingsfase waarin poederdeeltjes door de opgelegde druk dichter tot elkaar worden gebracht en een relaxatiefase die voor het grootste deel plaatsvindt na het verwijderen van deze druk. Deze fasen worden meestal gescheiden besproken, ondanks het feit dat de gebeurtenissen die plaatsvinden tijdens het verdichten van een poeder mede bepalend zijn voor het relaxatiegedrag van een tablet. Hoofdstuk 2 behandelt het tabletteergedrag van binaire mengsels bestaande uit natriumchloride dat beter bekend staat als keukenzout en verstijfseld aardappelzetmeel. De materialen zullen voor het overige deel van deze samenvatting worden aangeduid als zout respectievelijk zetmeel. Deze twee materialen werden uitgekozen, omdat ze zeer verschillen in verdichtings- en relaxatiegedrag. Ook zijn de krachten waarmee deeltjes met elkaar binden verschillend. Zout bindt met zoutbruggen en zetmeel voornamelijk met waterstofbruggen. Hierdoor is de interactie tussen zout en zetmeel zwak. Door het variren van de concentraties van beide stoffen

111

Samenvatting

in het mengsel wordt het mogelijk geacht om de veranderingen in tabletsterkte aan het verdichtings- en relaxatiegedrag van mengsels te koppelen. Het onderzoek van het compactiegedrag van binaire mengsels begint al bij het mengen van de twee materialen. In het begin is ervoor gekozen om met een gelijke deeltjesgrootte voor beide stoffen te beginnen. In dit onderzoek werden de mengsels beoordeeld op volumebasis. De tabletten werden geslagen met een constante tabletteersnelheid van 3 mms-1, omdat de tabletteersnelheid ook van invloed is op de verdichting en relaxatie van poeders. Echter, de tabletteersnelheid vormt geen onderdeel van dit onderzoek. Het productieproces van een tablet begint met het storten van het poeder in een matrijs. De deeltjes zullen zich eerst door een lichte druk van de stempels ten opzichte van elkaar gaan verplaatsen tot het moment dat er geen verdere herrangschikking mogelijk is. Verdere verdichting van het poederbed is alleen mogelijk door vervorming van de deeltjes. Daar zout en zetmeel bekend staan als visco-elastische materialen, zullen de deeltjes plastisch vervormen, dat wil zeggen dat de deeltjes alleen van vorm veranderen en niet in kleinere deeltjes zullen opbreken. De druk waarbij deeltjes plastisch vervormen heet de zwichtspanning. Zout heeft een hoge zwichtspanning en zetmeel een lage zwichtspanning. De zwichtspanningen van de mengsels vertonen bijna een lineaire relatie met de mengselsamenstelling. Dit betekent dat de zwichtspanning van een mengsel voorspeld kan worden vanuit de volumefracties van beide stoffen en de individuele zwichtspanningen. In dit hoofdstuk wordt aangetoond dat er kleine negatieve afwijkingen ten opzichte van deze lineaire relatie bestaan. Op elektronenmicroscopiefotos van dwarsdoorsneden van tabletten is te zien dat de initile deeltjesvorm van sommige zoutdeeltjes onveranderd is gebleven. Deze deeltjes zijn dus niet gezwicht tijdens het tabletteerproces. Dit wordt verklaard door het verschil in zwichtspanning tussen zout en zetmeel. Wanneer een hard zoutdeeltje omringd is door zachte zetmeeldeeltjes, zullen deze zetmeeldeeltjes om het zoutdeeltje heen vormen en het zoutdeeltje beschermen tegen de opgelegde druk. Dit verklaart de kleine negatieve afwijkingen. De relaxatie van de tabletten kan worden weergegeven als de porositeitexpansie wat het verschil is tussen de porositeit (volumefractie lucht) onder de maximale druk tijdens het tabletteerproces en de porositeit van het tablet na relaxatie. Zouttabletten hebben een lage porositeitexpansie en zetmeeltabletten een hoge porositeitexpansie. De tabletten geslagen van de mengsels hebben echter hogere porositeitexpansies dan was verwacht op basis van lineaire interpolatie. Het verschil werd gedefinieerd als de extra porositeitexpansie. Gezien het feit dat de tabletrelaxatie wordt geremd door

112

Samenvatting

binding tussen deeltjes, wordt deze extra porositeitexpansie beschouwd als het gevolg van lagere bindingskrachten tussen deeltjes van de verschillende materialen. Het uiteindelijke tablet met de bijbehorende tabletstructuur en tabletsterkte is het resultaat van de gebeurtenissen tijdens de verdichtings- en relaxatiefase. Veelal wordt gekeken naar het verschil in tabletsterkte van tabletten die allemaal geslagen zijn bij dezelfde compactiedruk. Het is echter bekend dat niet zozeer de compactiedruk of compactiesnelheid, maar nog meer de uiteindelijke tabletstructuur, inclusief de porositeit, die bepalend is voor de tabletsterkte. Om deze reden werd het verloop van de tabletsterkte bestudeerd bij een constante porositeit van 15%. De tabletten geslagen van de mengsels hadden lagere tabletsterktes dan was verwacht vanuit de volumefracties en materiaalsterktes bij 15% porositeit. Een afname van tabletsterkte door het mengen van verschillende materialen duid ook op een afname van de bindingskrachten tussen de deeltjes. Hoofdstuk 2 toont aan dat het verschil tussen de gemeten en verwachte tabletsterktes overeenkomt met de extra porositeitexpansie van deze zout-zetmeel mengsels. In hoofdstuk 3 wordt dieper ingegaan op de veranderingen in tabletstructuur van tabletten bestaande uit zout-zetmeel mengsels. Met behulp van een warmtebehandeling in de oven werden de zetmeeldeeltjes uit de tabletten gebrand. Dit is mogelijk door het verschil in de zelfontbrandingstemperatuur van zetmeel (300C) en de smelttemperatuur van zout (800C). Hierdoor ontstaat een tabletstructuur die alleen bestaat uit zoutdeeltjes. Door het meten van de grootte van de nieuwe porin (gaten lucht in de tabletstructuur) werd een indicatie verkregen van de deeltjesvorm van zetmeel in het tablet. Aangetoond kon worden dat de zetmeeldeeltjes extreem vervormd worden tot platte deeltjes, terwijl de deeltjesdiameter van zout ongeveer gelijk blijft. Dit kan wederom aan het verschil in hardheid tussen beide materialen worden toegeschreven. Een tabletstructuur bevat naast de vaste deeltjesstructuur ook een poriestructuur. Wanneer de poriegrootteverdelingen van tabletten bestaande uit de mengsels met een porositeit van 15% met elkaar vergeleken werden, kwam naar voren dat de kleine porin constant bleven en grotere porin veranderden in grootte en aantal. Vooral in de tabletten geslagen van de mengsels waren grote porin aanwezig. Op basis van de theorie over veranderingen van de poriegrootte in tabletten geslagen van viscoelastische materialen, kon er beredeneerd worden dat deze porin vooral het gevolg zijn van de tabletrelaxatie van mengsels. Om deze toename van de grote porin te kunnen weergeven in een enkele waarde, werd de gemiddelde poriediameter (d50 van het cumulatieve porievolume) voor elke mengselsamenstelling berekend. Daarnaast is de extra porositeitexpansie (hoofdstuk 2) ook het resultaat van het

113

Samenvatting

relaxatiegedrag van mengsels. Hoofdstuk 3 laat zien dat er een lineair verband bestaat tussen de extra porositeitexpansie en de gemiddelde poriediameters voor zout-zetmeel mengsels met gelijke initile deeltjesgrootte. Hoofdstuk 4 concentreert zich vooral op de sterkte van de structuren die gevormd worden door zoutdeeltjes in de verschillende tabletten. Een continue structuur van deeltjes van n materiaal wordt ook wel een matrix genoemd. De concentratie waarbij deze continue structuur verandert in losse clusters van deeltjes is de percolatiedrempel van dat materiaal in het tablet. Door het verwijderen van de zetmeeldeeltjes kan de sterkte van de zoutmatrix vergeleken worden met de initile tabletsterkte en kan de percolatiedrempel van zout worden bepaald. De percolatiedrempel van zetmeel in deze tabletten werd bepaald door middel van desintegratietests. De sterkte van de zoutmatrix na het verwijderen van de zetmeeldeeltjes komt overeen met de initile sterkte van de tabletten alleen als de zetmeeldeeltjes in het tablet aanwezig zijn als clusters. Dit bewijst dat de zoutmatrix volledig de tabletsterkte bepaalt van deze tabletten. Dit betekent ook dat het negatieve effect van de zetmeeldeeltjes op de tabletsterkte al wordt bewerkstelligd tijdens het compactieproces. Wanneer de zetmeeldeeltjes ook een continu netwerk in het tablet vormen, dan is de sterkte van de zoutmatrix lager dan de initile tabletsterkte. Als de zoutdeeltjes geen matrix meer vormen en de zetmeeldeeltjes worden verwijderd, vallen de tabletten uiteen en is de matrixsterkte van zout gelijk aan nul. De sterktes van de zoutmatrices in de tabletten bestaande uit de mengsels voor en na het uitbranden van het zetmeel werden vergeleken met de sterktes van de tabletten die werden geslagen van puur zout. De zoutmatrices in deze pure zouttabletten zijn veel sterker dan die in de tabletten van de mengsels. De tabletsterkte werd vervolgens gecorreleerd aan de toename van de grote porin die beschreven zijn in hoofdstuk 3. Porin kunnen zich gedragen als een wig; tijdens een sterktemeting wordt de spanning geconcentreerd aan het eind van de porie waardoor het tablet bij lagere krachten zal breken. Grotere porin veroorzaken in het algemeen een grotere verzwakking van de tabletsterkte. Vanuit dit standpunt werd de gemiddelde poriediameter gerelateerd aan de sterkte van de tabletten (tabletten bestaande uit de mengsels waarvan de zoutmatrix de sterkte bepaalt, de uitgebrande tabletten en de pure zouttabletten). Bij alle drie soorten tabletten bestaat er een directe relatie tussen de gemiddelde poriediameter en de tabletsterkte, waarbij de pure zouttabletten hogere tabletsterktes hebben dan de tabletten geslagen van mengsels. Naast de poriegrootte is ook de porievorm van belang voor de karakterisering van de tabletsterkte. Een relatief lange en/of scherp porie heeft een grotere capaciteit om spanning te concentreren en kan zodoende een grotere afname van de tabletsterkte realiseren. Om een porievorm te kunnen

114

Samenvatting

definiren, werden de porin beschreven als elliptische vormen. Hierdoor is het mogelijk om de porievorm te definiren als de verhouding tussen de lengte en diameter. Uit de berekeningen blijkt dat de porin in de tabletten geslagen van de mengsels relatief lang en/of scherp zijn. Door de vervorming van de zetmeeldeeltjes tot platte deeltjes en de opvolgende tabletrelaxatie worden grotere porin gegenereerd die een overeenkomende uitgerekte vorm aannemen. Om deze conclusies verder uit te werken worden in hoofdstuk 5 de effecten van veranderingen in deeltjesgrootte van zout op de sterkte en de percolatiedrempel van de zoutmatrix bestudeerd. Door het gebruik van kleinere deeltjes in de mengsels kunnen de zoutdeeltjes gemakkelijker een matrix vormen in de tabletstructuur. Hierdoor verschuift de percolatiedrempel van zout naar lagere concentraties. In tegenstelling tot de zoutdeeltjes vormen de zetmeeldeeltjes altijd een matrix bij een lage concentratie. Door de extreme vervorming van het zetmeel vormen deze deeltjes altijd een matrix bij een lage concentratie en is de percolatiedrempel onafhankelijk van de deeltjesgrootte van zout. Enerzijds zorgen kleinere zoutdeeltjes voor een toename van de tabletsterkte van tabletten bestaande uit puur zout en mengsels. Anderzijds veroorzaken zetmeeldeeltjes ook in de tabletten, waarvan de zoutmatrices met verschillende deeltjesdiameters de tabletsterkte bepalen, een toename van de gemiddelde poriediameter tijdens tabletrelaxatie. Deze grotere porin zijn de voornaamste reden voor de verzwakking van de tabletsterkte. Overeenkomend met hoofdstuk 4 wordt de unieke relatie bevestigd tussen gemiddelde poriediameter en tabletsterkte voor tabletten geslagen van mengsels. De relatie is onafhankelijk van de deeltjesgrootte van zout en de zoutconcentratie. Voor pure zouttabletten wordt een afwijkende relatie gevonden. Het verschil tussen beide unieke relaties wordt wederom verklaard door de porievorm. Het uiteindelijke doel van het onderzoek naar het compacteergedrag van binaire mengsels is het voorspellen van de tableteigenschappen uit de individuele materiaaleigenschappen. In hoofdstuk 6 wordt een model gentroduceerd dat de veranderingen in de tabletstructuur beschrijft aan de hand van de individuele materiaaleigenschappen en de invloeden hiervan op de eigenschappen van tabletten bestaande uit twee materialen. Het model is geschikt om tableteigenschappen (elasticiteitsmodulus en tabletsterkte) van tabletten zonder een fractie lucht te beschrijven. Dit resulteert in het kunnen beschrijven van een relatieve adhesiekracht tussen zout en zetmeel. In hoofdstuk 7 worden de conclusies beschreven aan de hand van de materiaaleigenschappen van beide stoffen en worden voorstellen gedaan voor verder onderzoek.

115

Samenvatting

116

Dankwoord

Dankwoord
Aan het begin van je AiO-periode denk je dat onderzoek op eigen houtje uitgevoerd kan worden. Het blijkt toch echter maar weer dat vele handen licht werk maken. Ik wil een aantal mensen bedanken die mijn onderzoek in goede banen hebben geleid. Ik zal waarschijnlijk verschillende mensen vergeten, maar ik hoop dat zij beseffen dat ze ook een positieve invloed op mijn periode als AiO hebben gehad. Ten eerste wil ik Erik Frijlink bedanken voor het aanbieden van deze onderzoeksplek en dat hij daarvoor bereid was de positie een half jaar vrij te houden. Het was ook altijd mogelijk om zijn kantoor binnen te lopen met een vraag of probleem. Erik, bedankt voor de inspiratie en de steun die ik heb gevoeld de afgelopen jaren. Gerad Bolhuis wil ik als co-promoter bedanken voor de uren van discussie. Het uittekenen van resultaten op het witte bord heeft zeker geholpen mijn ideen in de juiste volgorde te zetten. Jouw ervaring en contacten hebben mij zeer geholpen. Als laatste van de directe begeleidergroep wil ik Kees van der Voort Maarschalk bedanken voor de tijd die hij vrijgemaakt heeft tijdens zijn eigen werkzaamheden bij Organon. Met zijn kennis heeft hij ervoor gezorgd dat ik op het rechte pad bleef en de goede afslagen nam tijdens het onderzoek. Kees, bedankt voor de vele aanwijzingen tijdens het schrijven van artikelen. Ik moet toegeven dat ik veel geleerd heb van jou. I would like to thank the members of the reading committee, Jarkko Ketolainen, Herman Vromans and Hans Wesselingh, for the critical evaluation of the manuscript of this thesis. Furthermore, I would like to thank Jarkko Ketolainen and the other colleagues at the department of Pharmaceutical Technology and Biopharmacy of the University of Kuopio for the warm stay during the cold winter of 2001-2002. Klaas, ondanks de langdurige periodes van afwezigheid gedurende deze vier jaar, waren de tijden als kamergenoot zeer geslaagd. Hierin wil ik ook Anko en Wouter betrekken die met hun grappige verhalen de koffiepauzes zeer gezellig maakten. Ik heb ook veel geleerd van hun technische en fysische inzichten. Als CSI-er (Committed Subject Investigator) zijn je mede-AiOs zeer belangrijk. Niet alleen voor steun in het doen van onderzoek, maar vooral voor het uitwisselen van frustraties en de laatste roddels en voor de gezelligheid tijdens de lunch, etentjes en de uren op de tennisbaan. Dit geldt zowel voor de vroegere groep (Hans, Joke, Paul, Rob en Stefan) als voor de huidige groep AiOs (Anneke, Bastiaan, Dirk-Jan, Gerrit en Jonas).

117

Samenvatting

Ik heb het altijd plezierig gevonden om studenten te begeleiden. Naast het geven van de cursus Formuleringen was de begeleiding van Maaike, Yu San en Esther zeer verfrissend. De EIK (Eerstejaars Introductie Kommissie) leverde naast veel extra werk ook veel uren van plezier op. Een gemotiveerde groep studenten en AiOs die met elkaar zeer geslaagde introductiedagen en weekend hebben georganiseerd. Naast de overige collegae en personen in de werksfeer, zijn vrienden en familieleden van groot belang. De jaren waarin ik samen met mijn vrienden uit Oldeberkoop veel beleefd heb zal ik niet snel vergeten. Het voetballen in het weekend zorgde voor de nodige afleiding en lichamelijke gezondheid (ondanks het feit dat er wel een aantal blessures waren). Ondanks de veranderde situatie hoop ik dat er nog tijden komen dat we de leuke momenten kunnen herbeleven. Ook wil ik de vrienden bedanken die mijn studie- en AiO-tijd in Groningen tot een plezierige tijd hebben gemaakt. Speciaal wil ik mijn paranimfen, Mattijs Julsing en Michiel van der Wal, bedanken voor de gezellige studietijd en voetbal/Kolonistenavonden. De laatste paragraaf van dit dankwoord wil ik toewijden aan familie en mensen die dicht bij mij staan. Ensin min haluan kiitt minun tyttystvni perhett tuesta (Ten eerste wil ik de familie van mijn vriendin bedanken voor de steun). Mijn zussen, Geertje en Grietje, waardeer ik zeer en de dagen waarin we allemaal bij elkaar zijn, zijn ouderwets gezellig. Naar mijn ouders gaat een zeer grote dank uit. Ze hebben mij altijd gesteund en gestimuleerd in alles wat ik heb gedaan. Ik ben ze daarvoor altijd dankbaar. Emilia, the last words are for you. It has been almost four years ago that we met each other. A lot has happened since. You are not only my girlfriend, but also my best friend that supports and comforts me. There stands now a new period ahead of us. I am sure we will make it together.

118