Pharmaceutical Technology Division Department of Pharmacy University of Helsinki Finland

RHEOLOGICAL PROPERTIES OF PHARMACEUTICAL CREAMS CONTAINING SORBITAN FATTY ACID ESTER SURFACTANTS

Mirka Korhonen

ACADEMIC DISSERTATION

To be presented, with the permission of the Faculty of Science of the University of Helsinki, for public criticism in Auditorium 2041 of Biocentre Viikki (Viikinkaari 5E) on January 2nd, 2004, at 12 noon

Helsinki 2003

Supervisors:

Professor Jouko Yliruusi Pharmaceutical Technology Division Department of Pharmacy University of Helsinki Finland Professor Jouni Hirvonen Pharmaceutical Technology Division Department of Pharmacy University of Helsinki Finland

Reviewers:

Professor Kristiina Jrvinen Department of Pharmaceutics Faculty of Pharmacy University of Kuopio Finland Sari Kallioinen, Ph.D. University Pharmacy Helsinki Finland

Opponent:

Professor Liisa Turakka National Agency for Medicines Helsinki Finland

Mirka Korhonen 2003 ISBN 952-91-6585-4 (paperback) ISBN 952-10-1476-8 (PDF) Yliopistopaino Helsinki 2003 Finland

ABSTRACT

Korhonen, M., 2003. Rheological properties of pharmaceutical creams containing sorbitan fatty acid ester surfactants. ISBN 952-91-6585-4 (paperback) ISBN 952-10-1476-8 (PDF) The main purpose of the present study was to gain understanding about the rheological (elastic, viscoelastic and viscous) properties of pharmaceutical creams containing sorbitan fatty acid ester surfactants (sorbitan monolaurate, sorbitan monopalmitate, sorbitan monostearate, sorbitan monooleate and sorbitan trioleate). The study investigated the effects of an increase in the hydrocarbon chain length, the double bonded hydrocarbon chains and the concentration of the surfactant on the rheological properties of the sorbitan fatty acid ester surfactant containing creams. In addition, the effects of the volume of inner phase and the short-term storage on the rheological properties of the sorbitan fatty acid ester surfactant containing creams were investigated. The creams studied were either simple, three-component creams, or complex, multi-component creams. The rheological properties were determined with dynamic oscillation stress sweep and oscillation frequency sweep measurements, with static creep recovery measurements and with time-dependent viscosity measurements. An increase in the length of the saturated hydrocarbon chain of the surfactant increased the elastic nature of the creams. A double bond in the hydrocarbon chain of the surfactant decreased, and an increase in the number of double bonded hydrocarbon chains within the surfactant molecule increased the elastic nature of the creams. An increase in the concentration of the surfactant and an increase in the volume of the inner phase increased the elastic nature of the creams. Storage time decreased and storaging the creams at three different conditions levelled off the differences in the elastic properties of the creams. In the three-component cream formulations the rheological properties of the creams were concluded to be mainly due to the interfacial properties of the surfactants. In the multi-component cream formulations the rheological properties of the creams were concluded to be due, in addition to the surfactants used, to the multiphase structures of the creams. Determination of rheological properties, i.e. unrecoverable viscous, partly recoverable viscoelastic and recoverable elastic properties, gave valuable information about the structural and intermolecular properties of the creams. Rheological measurements can be regarded as sensitive tools for detecting structural changes in pharmaceutical creams and should be regarded as an integral part of the quality evaluation of pharmaceutical creams.

TABLE OF CONTENTS

TABLE OF CONTENTS ACKNOWLEDGEMENTS LIST OF ABBREVIATIONS 1. INTRODUCTION 2. THEORY 2.1 Theories of emulsification 2.2 Surfactant orientation in pharmaceutical creams 2.2.1 Orientation at interfaces 2.2.2 Micelle formation 2.2.3 Additional structures in pharmaceutical creams 2.3 Instability of pharmaceutical creams 2.4 Rheological properties of pharmaceutical creams 2.4.1 Viscosity 2.4.2 Elasticity 2.4.3 Viscoelasticity and viscoelastic models 2.4.4 Viscoelastic test methods 3. AIMS OF THE STUDY 4. EXPERIMENTAL 4.1 Materials 4.2 Methods 4.2.1 Cream preparation (I-IV) 4.2.2 Determination of rheological properties (I-IV) 4.2.3 Determination of interfacial tension (IV) 4.2.4 Determination of droplet size distribution (I-III) i

i iii iv 1 3 3 5 5 7 11 14 16 16 20 20 22 27 28 28 29 29 30 31 31

4.2.5 Determination of conductivity (I, II) 5. RESULTS AND DISCUSSION 5.1 Effect of the length of the saturated hydrocarbon chain on the rheological properties of the creams (III, IV) 5.2 Effect of the double bonds in the hydrocarbon chains on the rheological properties of the creams (I, IV) 5.3 Effect of the concentration of the surfactant on the rheological properties of the creams (III) 5.4 Effect of the volume of the inner phase on the rheological properties of the creams (III) 5.5 Effect of storaging on the rheological properties of the creams (I, II) 6. CONCLUSIONS REFERENCES

31 32

32

37

40

44 46 51 52

ii

ACKNOWLEDGEMENTS

This study was carried out at the Pharmaceutical Development Department of Orion Corporation Orion Pharma, Turku, and at the Pharmaceutical Technology Division, Department of Pharmacy, University of Helsinki, during the years 1996-2003. I express my deepest gratitude to Professor Jouko Yliruusi for his valuable instructions and enthusiasm, and for giving me the opportunity to do my studies at my own pace. My sincere appreciation also goes to Professor Jouni Hirvonen for his valuable advises and support during this study. My respectful thanks go to Professor Kristiina Jrvinen and Sari Kallioinen, Ph.D., the reviewers of this thesis, for their critical evaluation and constructive criticism. I am thankful to Juha Kiesvaara, Ph.D., for giving me the opportunity to continue my studies at the Pharmaceutical Development Department of Orion Corporation Orion Pharma, Turku. I also owe my warmest thanks to my boss Esko Taskila, M.Sc., for his support and understanding during the years of my studies. I owe my special thanks to my co-authors, Heikki Niskanen, M.Sc., for introducing for me the fascinating world of rheology, Johanna Lehtonen, M.Sc., and Laura Yrjninen, M.Sc., for their help with rheological measurements, Leena Peltonen, Ph.D., for her help with the interfacial tension measurements, and Osmo Antikainen, Ph.D., for his collaboration. I also thank Docent Leena Helln for her support and enthusiasm during my studies. I thank Sari Kuivamki, M.A., for revising the English manuscript of this thesis and Ari Lempiinen for his technical assistance. Finally, I thank my family and my friends for their everlasting support and enjoyable and irreplaceable moments outside my studies and my work.

Kaarina, November 2003

iii

LIST OF ABBREVIATIONS

Acmc c cmc
&

area per molecule just below the cmc concentration critical micelle concentration phase angle viscosity interfacial tension shear rate strain surface excess rigidity (shear, elastic) modulus storage modulus loss modulus complex modulus compliance Avogadros number angular deflection gas constant shear stress absolute temperature loss tangent

G G G G* J N R T tan

iv

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following original papers, which are referred to in the text by Roman numerals I-IV I Korhonen, M., Niskanen, H., Kiesvaara, J., Yliruusi, J., 2000. Determination of optimal combination of surfactants in creams using rheology measurements. Int. J. Pharm. 197, 143-151. Korhonen, M., Helln, L., Hirvonen, J., Yliruusi, J., 2001. Rheological properties of creams with four different surfactant combinations effect of storage time and conditions. Int. J. Pharm. 221, 187-196. Korhonen, M., Lehtonen, J., Helln, L., Hirvonen, J., Yliruusi, J., 2002. Rheological properties of three component creams containing sorbitan monoesters as surfactant. Int. J. Pharm. 247, 103-114. Korhonen, M., Hirvonen, J., Peltonen, L., Antikainen, O., Yrjninen, L., Yliruusi, J., 2003. Formation and characterization of three-component sorbitan monoester surfactant, oil and water creams. Int. J. Pharm., in press.

II

III

IV

1. INTRODUCTION

Emulsions are thermodynamically unstable colloidal systems containing two immiscible liquids, oil and water, one of which is finely dispersed into the other. To form a stable emulsion, also a third component, surfactant, is needed. With both polar and non-polar regions at the same molecule, surfactant molecule settles at the interface of oil and water and decreases the interfacial free energy, interfacial tension, between them (Cosgrove, 1963; Myers, 1988). The film structures of surfactant molecules at the interfaces form the basis of homogenous and stable emulsions. Emulsions can be divided into either oil-in-water (o/w) or water-in-oil (w/o) emulsions. In o/w emulsion oil droplets are dispersed throughout the aqueous phase, while in w/o emulsion aqueous droplets are dispersed throughout the oil phase (Fig. 1). In addition to these simple emulsions, also multiple emulsions can be prepared. These can be either o/w/o or w/o/w emulsions, respectively.

Figure 1. Structural principle of o/w and w/o emulsions (reproduced from Stokes and Evans, 1997). Surfactant molecules gather at the interfaces so that the hydrophilic head group orientate towards the water phase and the hydrophobic tail towards the oil phase.

Emulsions that are semisolid in consistency, with a high apparent viscosity, are called creams (Billany, 1988). Creams are prepared for external use. Creams contain excess surfactant over that required to form a monomolecular surfactant film at the interfaces (Eccleston, 1986a). This excess interacts with other components either at droplet interfaces or in the bulk phase to produce complex, multiphase structures. These complex multiphase structures are essential to form stable pharmaceutical creams for extended periods of time. Determination of rheological properties - unrecoverable viscous, partly recoverable viscoelastic and recoverable elastic properties - gives reliable and versatile information about the structural properties of the cream. From these, viscoelastic properties are the most important rheological properties for pharmaceutical creams. Changes in the rheological properties may signify instability (Zografi, 1982) and provide qualitative and quantitative information about the structural, intermolecular properties of the cream. It is generally concluded that the viscoelastic nature of a cream is a good indicator of its physical stability (Zografi, 1982; Frster and Herrington, 1997; Gasperlin et al., 1998). Sorbitan fatty acid ester surfactants are widely used non-ionic surfactants in pharmaceutical creams. Although the rheological properties of sorbitan fatty acid ester surfactant containing creams have been studied, sorbitan fatty acid ester surfactants are usually used as cosurfactants in these study creams (Boyd et al., 1972; Groves and de Galindez, 1976; Kawashima et al., 1991,1992; Lashmar and Beesley, 1993; Carlotti et al., 1995; Lee et al., 1997; Opawale and Burgess, 1998a; Bjerregaard et al., 2001; Jiao and Burgess, 2003). In addition, the rheological properties are usually studied only with viscosity measurements (Groves and de Galindez, 1976; Kawashima et al., 1991,1992; Carlotti et al., 1995; Jiao and Burgess, 2003). There is a lack of determinations of viscoelastic properties of pharmaceutical creams containing sorbitan fatty acid ester surfactants. In addition, there is a lack of studies, with minimal variables, on how the differences in the molecular structures of single, not combined, sorbitan fatty acid ester surfactants affect the rheological properties of pharmaceutical creams.

2. THEORY

2.1 Theories of emulsification

The preparation of a pharmaceutical cream requires the formation of a very large interfacial area between two immiscible phases, oil and water. Surfactants facilitate the preparation of an emulsified system by lowering the free energy in the system and by reducing the work required to generate new interfaces (Myers, 1988). There are several theories that have been implemented to explain the process of emulsification. The next chapters describe a few theories of emulsification by non-ionic surfactants. Non-ionic surfactants form the most widely used group of surfactants. Since non-ionic surfactants do not carry a charged group in their hydrophilic head group, they are not as sensitive to the influences of present components and surrounding conditions as ionic surfactants. In addition, in topical formulations non-ionic surfactants cause less skin irritation than ionic surfactants do.

Theories of surface tension and interfacial behaviour

Surface tension is a result of attractive cohesive forces between molecules adjacent at the surface and below (Martin, 1993). Tension between two immiscible liquids is called interfacial tension. According to the surface tension theory, an emulsion may be formed of two immiscible liquids if an agent that lowers appreciably the interfacial tension is added to the system (Cosgrove, 1963). A surfactant, due to the molecular character, settles at the interfaces of oil and water. The hydrophilic portion of surfactants orientates towards the aqueous phase and the hydrophobic portion towards the oil phase. The relative difference in the size and strength of the polar and non-polar groups determines the emulsification capability of the surfactant. Generally, if the hydrophilic properties of the surfactant are slightly more dominant than the hydrophobic properties, the molecule will orientate at an oil-water interface so that the hydrophobic portion is forced to the center, and the resulting emulsion is of an o/w type (Cosgrove, 1963). If the hydrophobic character of the surfactant is slightly more dominant than the hydrophilic character, w/o emulsions are typical. The balance between the hydrophilic and hydrophobic characters within the surfactant molecule is presented with the HLB, hydrophile-lipophile-balance, value. This balance was devised by Griffin in 1949. 3

According to the HLB system, the higher the HLB value of the surfactant, the more hydrophilic the surfactant is. Likewise, the lower the HLB value of the surfactant, the more hydrophobic the surfactant is. It has been stated that the HLB values 3-8 enhance the formation of w/o emulsions and 8-16 the formation of o/w emulsions (Martin, 1993).

Adsorbed-interfacial film theories

According to Bancroft et al., surfactants form a film around the dispersed droplet, thus preventing it from irreversible fusion, coalescence, with other dispersed droplets (Navarre, 1941). In 1913 Bancroft related the solubility of the surfactant to the type of the emulsion formed; the phase in which the surfactant is more soluble will be the continuous or external phase (Becher, 1957). This generalization is known as Bancrofts rule. A more elaborate form of this rule was stated by Bancroft and Tucker in 1927 (Becher, 1957). According to Bancroft and Tucker, the existence of an interfacial film requires the presence of two interfacial tensions (i.e. one on each side of the interfacial film) (Becher, 1957). The type of the formed emulsion depends on the relative strength of the surface tension on each side of the film (Cosgrove, 1963). The film curves in the direction of the higher interfacial tension; the disperse phase is thus on the side of the film with the higher interfacial tension.

Harkins oriented wedge theory

According to the oriented wedge theory, the difference in the relative size of the polar and non-polar groups of the surfactant explains the type of the formed emulsion. The group with the larger cross sectional area will be oriented outside of the droplet (Autian, 1966). For example, the greater cross sectional area of the polar rather than the non-polar group will produce an o/w emulsion.

Viscosity theory

Viscosity aids the emulsification process by increasing the resistance to droplet coalescence. Viscosity of an emulsion may have an effect on droplet size and on the stability of the 4

emulsion by slowing down the diffusion and Brownian movement and, hence, retarding the coalescence of droplets (Dunker, 1960; Cooper and Gunn, 1950). However, emulsions with high viscosity may also have poor stability and emulsions with low viscosity may have good stability. According to Stokes law (Martin, 1993), the stability of the emulsion may also be improved by reducing droplet size and droplet size distribution and by minimizing the density difference between the inner and continuous phases in addition to increasing the viscosity of the continuous phase.

Mixed surfactant theory

According to mixed surfactant theory, the most favourable surfactant combination develops if a hydrophilic surfactant is capable of forming a complex with a hydrophobic surfactant. Due to the molecular complex of the surfactant molecules formed, the interfacial film withstands greater pressures than either component alone (Dunker, 1960). Mixtures produce a stronger film with greater resistance to rupture and favour a more stable emulsion, i.e. the emulsion droplets are less liable to coalescence.

2.2 Surfactant orientation in pharmaceutical creams

2.2.1 Orientation at interfaces

Since amphiphilic surfactant molecules have both polar and non-polar groups in the same molecule, they minimize unfavourable solvophobic (solvent-hating) interactions by concentrating spontaneously at phase interfaces (Stokes and Evans, 1997). They preferentially adsorb at available interfaces by replacing the energy rich bulk phase molecules and thus reduce the free energy of the system. Interface can be described as the boundary between at least two immiscible phases. The boundaries between the phases are of primary importance in determining the characteristics and behaviour of the system as a whole (Myers, 1988). The stability of creams depends on the ability to control and manipulate phase boundaries and interfacial interactions. At the liquid-liquid interfaces the surfactant molecules will normally form monolayers with various molecular packing densities ranging from relatively loosely

packed arrangements to close-packed arrangements (Fig. 2) (Myers, 1988). These arrangements are primarily dependent on the molecular structure of the surfactant. In the liquid-expanded state arrangements (Fig. 2a), the adsorbed molecules are more or less perpendicular to the interface (Myers, 1988). There are cohesive forces between the hydrocarbon chains, although they are in a highly disordered state and orientate horizontally towards the interface (Krog, 1975). In horizontal orientation the molecules require a lot of space, leading to looser packing and weaker chain-chain interactions between the molecules at the interfaces. For example, the polar affinity of the hydrocarbon chain, due to the polar group in the hydrocarbon chain or the shortness of the hydrocarbon chain, favours the orientation to the expanded horizontal state (Adamson, 1960; Rakshit et al., 1981). Also a double bond in the hydrocarbon chain decreases the hydrophobic chain-chain interactions (Feher et al., 1977) and loosens the packing of surfactants at the interfaces (Carlotti, 1995). Looser packing may cause stability problems for creams depending on the cream formulation as a whole.

(a) Moderately close-packed with significant chain mobility (liquid-expanded).

(b) Close-packed with tilted orientation and reduced chain mobility (liquid-condensed).

(c) Close-packed with essentially vertical orientation and very limited chain mobility (condensed solid). Figure 2. General forms of monomolecular surfactant films at interfaces (Myers, 1988).

In condensed films the surfactant molecules are very close-packed as the polar groups are oriented towards the water phase and the hydrocarbon chains are oriented vertically towards the oil phase (Krog, 1975). In the liquid-condensed state (Fig. 2b), the hydrocarbon chains are able to oscillate around their axis, but they have a very limited degree of mobility with respect to their position at the interface. In the condensed solid state (Fig. 2c), the surfactant molecules are close-packed with essentially vertical orientation to the interface (Myers, 1988). In the condensed solid state, the surfactants act like solid materials. For example, the increasing hydrocarbon chain length promotes the tendency towards the formation of condensed monolayers (Rakshit et al., 1981). Also other components in the cream formulation affect the packing of interfacial molecules. It is concluded that the shorter the hydrocarbon chain of the used oil, as compared to the length of the hydrocarbon chain of the used surfactant, the easier it is for the oil molecules to penetrate between the surfactant molecules (Stokes and Evans, 1997). This increases the hydrocarbon volume at the interfaces and condenses the molecular packing at the interfaces. The condensed packing of surfactant molecules enhances the formation of stable creams due to the more ordered structure, and thus greater stability, of the interfaces.

2.2.2 Micelle formation

Micelles are self-aggregated surfactant molecules (Kayes, 1988). The primary reason for micelle formation is to minimize the free energy in the system. After the interfacial saturation, micelle formation is the following step for the surfactant molecules to reach an energetically favourable state. The surfactant concentration at which single surfactant molecules start to self-aggregate at the bulk phase, is called critical micelle concentration (cmc). At cmc the solvophobic portions of surfactant molecules segregate from the solvent (Fig. 3). For example, in o/w creams the hydrophobic parts of surfactant molecules orientate inside the micelle to escape from the unfavourable polar aqueous environment.

Figure 3. Spherical micelles in polar medium (left) and in non-polar medium (right) (Stokes and Evans, 1997). In hydrophilic medium, the hydrophobic parts of surfactant molecules orientate inside the micelle and vice versa.

Cmc can be extrapolated at the intersection of two linear parts of the interfacial tension vs. log surfactant concentration plots (Fig. 4) (Attwood and Florence, 1983). In the plot, the interfacial tension shows an initial gradual decrease with the increasing concentration of the surfactant and eventually, before reaching the cmc, it becomes a linear function of the logarithm of concentration (Fig. 4). At the beginning of the linear portion, the interfacial saturation has been reached, i.e. the surface excess (number of molecules per surface area) of the surfactant does not increase further (Schott, 1980). However, the interfacial tension continues to decrease linearly with the increasing bulk concentration of the surfactant until the cmc is reached. Along the linear part, the surface excess is constant.

Figure 4. Schematic plot of surface or interfacial tension () versus logarithm of the surfactant concentration (c) (Attwood and Florence, 1983). Critical micelle concentration (cmc) is the intersection point of these curves. 8

The Gibbs equation expresses the equilibrium between the surfactant molecules at the surface or interface and those in the bulk solution (Attwood and Florence, 1983). The surface excess, , can be calculated by means of the Gibbs adsorption isotherm

= - (1 / R T) (d / d ln c) ,

[1]

where is the surface excess concentration (number of molecules per surface area), R is the gas constant, T the absolute temperature, the interfacial tension, and c the concentration of the surfactant. From the surface excess, the area per molecule just below the cmc (A cmc) can be calculated according to

Acmc = 1 / N,

[2]

where N is the Avogadros number. Spherical micelles have a characteristic size, and thus the increasing surfactant concentration increases the number of spherical micelles, not the size of micelles (Stokes and Evans, 1997). With the increasing concentration of surfactant molecules, the concentration of spherical micelles becomes so great that intermicellar interactions become important. If the amount of medium is insufficient to fill the spaces between the spherical or elongated micelles, a more ordered structuring of the solution occurs (Fig. 5) (Attwood and Florence, 1983).

Figure 5. A cylindrical micelle in polar medium (left) and a bilayer micelle in non-polar medium (right) (Stokes and Evans, 1997).

The molecular structure of the surfactant has an effect on the shape of the formed micelles. For example, it has been observed that simple straight-chain surfactants form lamellar phases if the hydrocarbon chain is not too short or if the hydrophilic part of the surfactant is not too large (Mollet and Grubenmann, 2001). In addition, it has been observed that non-ionic amphiphiles with two hydrocarbon chains of 1020 carbon atoms in length form lamellar structures in aqueous media. Due to the physical attraction to each other, the ordered amphiphilic aggregates can change their size and shape significantly in response to small changes in concentration, pH, temperature and pressure (Stokes and Evans, 1997).

10

2.2.3 Additional structures in pharmaceutical creams

When the interfaces are saturated and it is not energetically favourable to form simple micelles, surfactant molecules start to form more complex micellar structures to reduce the free energy in the system. The formed structures can be very complex and their types depend, in addition to the used surfactant, also on the other components in the cream, and their concentrations. Fig. 6 shows a ternary phase diagram of surfactant microstructures in a hypothetical three-component system. In the figure, between the one-phase areas (in white), there are concentration areas where two (lined) or more (in grey) of the presented microstructures appear. For example, as the surfactant concentration increases, structures can change from spherical micelles to cubic liquid crystals, and from cylindrical micelles to hexagonal liquid crystals (Fig. 6). Hexagonal and lamellar microstructures are the two main types of mesophases (Attwood and Florence, 1983). The hexagonal microstructure, named also the middle phase, is characterized by a hexagonal array of indefinitely long, mutually parallel rods. In the lamellar microstructure, named also the neat phase, the surfactant molecules are arranged in hydrophilic and hydrophobic layers. Like in the spherical micelles, also in these complex microstructures the solvophobic portions of the surfactant molecules are shielded from the unpleasant solvent environment.

11

Figure 6. A schematic phase diagram of a surfactant, oil and water system (Davis, 1994). The maximum concentration of a component is at the corner of the component name. Outside the composition triangle are presented the surfactant microstructures in various phases.

In commercial pharmaceutical creams, there are always more components than just the surfactant, oil and water. Commercial pharmaceutical creams are complex polydispersed systems containing several surfactants, which complement the properties of each other (Eccleston, 1990). As the number of components in a cream is higher, the internal structure of the cream is more complex. As the concentration of the surfactant is increased, the excess of the monomolecular interfacial films interacts with other components either at the interfaces or in the bulk phases to produce complex, multiphase formulations (Eccleston, 1986a). Although the surfactant concentration is higher at the oil-water interface than in either of the bulk phases, most of the surfactant molecules are in the water phase (hydrophilic surfactants), or in the oil phase (hydrophobic surfactants) (Weiner, 1986). Surfactant molecules are in a continuous movement partitioning between the bulk phases and the interface, and changing their orientation at the interface. 12

Creams stabilized with surfactant mixtures (surfactants with fatty alcohols) have been concluded to form four-phase creams (Junginger, 1984a,b; Eccleston, 1986b, 1990). In these four-phase creams the dominant structural elements are the hydrophilic and lipophilic gel phases (Fig. 7).

Figure 7. Gel structures of a four-phase o/w cream stabilized with a surfactant mixture; a = mixed crystal bilayer, b = interlamellary fixed water layer, a + b = hydrophilic gel phase, c = lipophilic gel phase, d = bulk water phase, e = lipophilic components (dispersed phase) (Junginger, 1984a).

The hydrophilic gel phases have, depending on the hydrophilic characteristics of their polar groups, a very strong swelling capacity, which stiffens the structure of the cream. In o/w creams the hydrophilic gel phase forms the continuous phase of the system (Fig. 7). It is important for the stability of o/w creams that a dynamic equilibrium is maintained between the fixed water in the interlamellary hydrophilic gel phase and the bulk water phase. The lipophilic gel phase immobilizes the dispersed oil phase (Fig. 7). Based on these phenomena, the aggregation and irreversible fusion, coalescence, of droplets are inhibited, assuming that the structure of the cream remains stable during storage.

13

2.3 Instability of pharmaceutical creams

As soon as a cream has been manufactured, time- and temperature-dependent separation processes occur (Rieger, 1986). These processes can be enhanced by both the creams internal structure and the surrounding conditions. As creams are thermodynamically unstable systems, they have a tendency to revert back to the original two-phase system with a minimum interfacial area (Block, 1989). As the free energy of the interface is the driving force for the irreversible fusion of droplets, creams can be stabilized by the inclusion of a surfactant in the system, as it concentrates at the oil-water interfaces (Attwood and Florence, 1983). During the destabilization, a cream structure undergoes several consecutive and parallel steps before the final stage of separated layers is reached (Fig. 8) (Sjblom, 1996). In the first step of instability, droplets move due to diffusion or stirring, and if the resistance between the droplets is not effective enough, they become aggregated to each other; flocculation has taken place (Fig. 8). Flocculation refers to the attachment of individual emulsion droplets to form flocs or loose assemblies, in which the identity of each is maintained (Myers, 1988). In flocculated systems the single droplets are replaced by twins (or multiples) separated by a thin surfactant film (Sjblom, 1996).

Figure 8. Examples of the steps of instability (Sjblom, 1996). The destabilization of an emulsified system passes through several stages, for example flocculation, coalescence and creaming, before the final separation occurs.

14

Flocculation can be a reversible process; it can be overcome by an input of much less energy than that required in the original emulsification process (Myers, 1988). In flocculation the thickness of the surfactant film at the interphases is reduced (Sjblom, 1996). When the critical thickness is overcome, the film bursts and two droplets unite to form a single droplet; coalescence has occurred (Fig. 8). The phenomenon when large droplets grow at the expense of smaller ones is called Oswald ripening (Eccleston, 1986a). In parallel with flocculation and coalescence, the droplets rise (creaming) or sink (sedimentation) through the medium due to the differences in the density of the dispersed and continuous phases (Fig. 8). In addition to the presented physical signs of instability, also phase inversion, microbiological contamination and chemical decomposition of creams can occur (Weiner, 1986). The stability of the interfacial film is strongly dependent on the adsorption-desorption kinetics of the surfactant, solubility, and interfacial rheological properties, such as elasticity and viscosity (Sjblom, 1996). A low rate of failure of the adsorbed film at the droplet interfaces is a primary factor in the stabilization of an emulsion system. Such phenomena are related to the condensity and elasticity of the interfacial film. It is concluded that a condensed and rigid interfacial film opposes the film thinning and rupture of this film (Eccleston, 1986a). There is a positive correlation between the interfacial elasticity and emulsion stability (Opawale and Burgess, 1998a,b; Kanouni et al., 2002). A perfectly stable emulsion would maintain the same number and size distribution of dispersed droplets per unit of volume or weight of the continuous phase (Garrett, 1965). The consistency of the continuous phase also assists in the prevention of flocculation and close encounters of dispersed droplets (Eccleston, 1986a). Although stability is a relative term, the degree of stability can be assessed by observing the rate of change in the physical properties of the cream (Attwood and Florence, 1983). Changes in rheological characteristics represent important early warnings of impending failure of the product (Rieger, 1991). Changes in the rheological properties have been determined both with viscosity measurements (Sherman, 1964; Lashmar and Beesley, 1993; Kallioinen et al., 1994; Tamburic et al., 1996; Gaspar and Maia Campos, 2003) and with viscoelasticity measurements (Radebaugh and Simonelli, 1983; Terrisse et al., 1993; Tamburic et al., 1996; Gasperlin et al., 1998). In these studies, for example, a decrease in consistency, changes in time-dependent viscous behaviour and changes in the ratio between viscous and elastic properties have been concluded to be signs of instability.

15

2.4 Rheological properties of pharmaceutical creams

Rheology means the study of the deformation and flow of matter (Barnes et al., 1989). The deformation will depend on the properties of the material (Tschoegl, 1989). Which property dominates, and what the values of the parameters are, depend on the stress and the duration of stress application (Barnes et al., 1989). Thus, a given material can behave like a solid or like a liquid, depending on the time scale of the deformation process. If the experiment is performed relatively slowly, the sample appears to be viscous rather than elastic; if the experiment is performed relatively fast, it appears to be elastic rather than viscous (Barnes et al., 1989). Shear stress causes strain in solids and rate of strain in liquids; solids deform and liquids flow (Schramm, 1994). Due to stress application, rearrangements take place inside the material. In a purely viscous material, all the energy required to produce the deformation is dissipated as heat. On the contrary, in a purely elastic material, all the energy required to produce the deformation is stored.

2.4.1 Viscosity

Viscosity is an expression of the resistance of fluid to flow; the higher the viscosity, the greater the resistance (Martin, 1993). Viscosity is a material property, which is independent of geometry (Tschoegl, 1989).

Newtonian behaviour

The basic law describing the flow behaviour of an ideal, Newtonian, liquid is

& =

[3]

& where is the shear stress, the viscosity and the shear rate (Schramm, 1994). In ideal

viscous liquids the viscosity is constant and independent of the applied stress (Fig. 9). The 16

rate of flow is directly related to the applied stress. Newtonian liquids start to flow when a stress is applied, and deformation stops instantly when the stress is removed (Morrison, 2001). In Newtonian liquids the viscosity is constant with respect to the time of shearing and it does not change in the re-testing situation (Barnes et al., 1989). Viscosity is only dependent on temperature (Briceno, 2000). Homogenous, low molecular weight liquids are examples of Newtonian liquids.

Figure 9. Typical rheograms illustrating Newtonian, pseudoplastic, dilatant and plastic flows (Barry, 1983).

Non-Newtonian behaviour, time-independent

With non-Newtonian liquids the viscosity is no longer dependent on the temperature alone, but also on many other factors, such as shear stress, time and measuring geometry (Bricano, 2000). A given shear stress will bring about a given shear rate (or the other way around), and the viscosity for this given shear stress is

& =/ .

[4]

The behaviour occurring when the consistency of the material decreases with the increasing rate of shear is called shear-thinning, and a material that obeys this behaviour is called pseudoplastic (Fig. 9) (Barry, 1983). The flow curve of pseudoplastic materials starts at the origin and the plot is non-linear. The curved rheogram results from a shearing action of longchain molecules or other arranged structures (Martin, 1993). As the shearing stress is 17

increased, the normally disarranged molecules start to align their long axes in the direction of flow; the orientation reduces the internal resistance of the material, and the release of solvent associated with the molecules lowers the concentration and the size of the dispersed molecules (Martin, 1993). Some materials show the opposite effect, shear-thickening, exhibiting higher viscosity when forced to flow at high rates (Morrison, 2001). Dilatant materials increase their resistance to flow as the shear rate is increased (Fig. 9) (Barry, 1983). Dilatant flow may be a result of dispersions containing a high concentration ( 50%) of small, deflocculated particles (Marriott, 1988). At rest the particles are closely packed with the interparticle volume being at the minimum (Martin, 1993). The amount of vehicle is sufficient to fill this volume and lubricate the particles to move relative to one another at low shear rates. As the shear stress is increased, the bulk expands or dilates and the particles, in an attempt to move quickly past each other, take on an open form of packing. This leads to a significant increase in the interparticulate void volume. At some point, the amount of vehicle becomes insufficient to fill the increased voids between the particles. The resistance to flow increases, because the fluid can no longer completely lubricate the particles. When an initial finite force is needed before any rheological flow can start, yield value occurs (Wood, 1986). Below the yield value, the material behaves essentially as an elastic solid (Barry, 1983). The yield value is due to the contacts between the adjacent particles, which must be broken before any flow can occur (Martin, 1993). Yield value is connected to flocculation; the more flocculated the system, the greater the yield value will be. Rheological behaviour that contains a yield value is called plastic flow, and the materials which exhibit it are called Bingham bodies (Fig. 9). At stresses higher than the yield value, the Bingham bodies flow with a constant viscosity, like Newtonian fluids. Also other materials, such as pseudoplastic and dilatant materials, can have a yield value.

Non-Newtonian behaviour, time-dependent

If the down curve (the shear stress is reduced after the desired maximum stress is reached) is not identical with the superimposed up curve (the shear stress is increased), time-dependent changes occur in the material. The measured viscosity can either increase or decrease with the time of shearing (Barnes et al., 1989). Changes in the material caused by the shearing process can be reversible or irreversible. The time-dependent changes of the system can be measured with an area of hysteresis loop (Fig. 10) (Martin, 1993). The hysteresis loop shows the effect of shearing history on the viscosity of the material (Marriott, 1988). 18

Thixotropic material becomes more mobile with the time of stirring (Fig. 10). In thixotropy the viscosity is gradually decreased under the shear stress and the structure is gradually recovered, when the stress is removed (Barnes et al., 1989). After the stirring is stopped, the system resets to the same consistency as it was before stirring, after a characteristic time lag for recovery (Barry, 1983).

Figure 10. Rheograms depicting time-dependent non-Newtonian behaviours (Barry, 1983). The up curve of hysteresis loop denotes the shear stress/shear rate relationship obtained with the increasing shear. The down curve of hysteresis loop represents the values derived by the subsequently decreasing shear.

Thixotropic materials usually contain asymmetric particles that through numerous points of contact set up a loose three-dimensional network throughout the sample (Martin, 1993). At rest, the structure confers some degree of rigidity. As the shear is applied and the flow starts, the structure begins to break down and the particles become aligned. The material undergoes a gel-to-sol transformation and exhibits shear-thinning. Upon removal of stress, the structure starts to reform. Shearing can also cause irreversible structural breakdown, changing the viscosity and yield value of the sample (Bricero, 2000). This behaviour can be referred to as shear destruction rather than thixotropy (Marriott, 1988). Anti-thixotropy is the opposite property to thixotropy. In anti-thixotropy the resistance to flow increases with the increased time of shear (Martin, 1993). Anti-thixotropy arises from the combination of dilatant and thixotropic behaviours (Fig. 10) (Barry, 1983). Like in the thixotropic materials, also in the anti-thixotropic materials the structure recovers after the stress application (Barnes et al., 1989).

19

2.4.2 Elasticity

Elastic deformations are reversible (Tschoegl, 1989). The applied stress causes instantaneous deformation; once the deformed state is reached there is no further movement and the deformed state remains as long as the stress is applied (Barnes et al., 1989). After the applied stress is removed, total recovery occurs. For a purely elastic material, rigidity modulus, G, is a material property

G = / ,

[5]

where is the shear stress and is the strain (Barnes et al., 1989). The proportionality between the stress and strain expressed by Equation [5] is consequently called Hookes law, and a material obeying this rheological equation of state is referred to as a Hookean solid. For the elastic materials the stress generated is directly proportional to the strain, that is, to the deformation (Morrison, 2001).

2.4.3 Viscoelasticity and viscoelastic models

Hookes and Newtons laws are linear laws, which assume direct proportionality between the stress and strain, or the rate of strain, whatever the stress (Barnes et al., 1989). The term viscoelasticity is used to describe behaviour which falls between the classical extremes of elastic response by the Hookean solids and the Newtonian liquids. Viscoelastic materials have simultaneously elastic and viscous properties. It has been concluded that viscoelasticity, shear-thinning and the presence of a yield value are properties of optimal pharmaceutical creams (Frster and Herrington, 1997); the creams remain consistent during storage and are easy to apply. In principle, all real materials are viscoelastic. Some energy may always be stored during the deformation of a material (Tschoegl, 1989), and there is always some unrecoverable deformation during long-term stress. Which property dominates, and what the values of the parameters are, depend on the stress and the duration of stress application (Barnes et al., 1989).

20

In the linear viscoelastic region the ratio of stress to strain is a function of time alone (Kobayashi et al., 1982). In the linear viscoelastic behaviour, doubling the stress will double the strain. Linear viscoelastic behaviour can be presented with mechanical models. These mechanical models consist of springs and dashpots arranged in parallel and/or in series. In these mechanical models the Newtonian flow is represented by a dashpot, an element in which the force is proportional to the rate of extension (Fig. 11a), and the Hookean deformation is represented by a spring, an element in which the force is proportional to the extension (Fig. 11b) (Barnes et al., 1989). A spring with recovering properties stores energy; a dashpot with non-recovering properties dissipates energy.

(a)

(b)

(c)

(d)

Figure 11. Dashpot (a) and spring (b) elements used in mechanical viscoelastic models. In Maxwell model (c) the dashpot and spring are connected in series. In Voigt model (d) the dashpot and spring are connected in parallel (Martin, 1993).

Maxwell model characterizes viscoelastic liquids (Schramm, 1994). In the model the spring and the dashpot are connected in series (Fig. 11c). In the Maxwell model the shear stresses in both the elements are always identical and the deformations are additive (Schramm, 1994). When a stress is applied, this model reacts first with an instantaneous increase in the strain due to the elastic response of the spring. At a later phase the fluid shows a viscous response in a dashpot, i.e. the deformation is unlimited as long as the stress is applied. After the removal of stress, the drop in strain is related to the release of stress in the spring, while the remaining permanent strain is equivalent to the amount of viscous flow in the dashpot. Voigt model represents a viscoelastic solid (Schramm, 1994). The Voigt model results from a parallel combination of a spring and a dashpot (Fig. 11d). A spring and a dashpot in parallel represent a material whose response to an applied stress is not instantaneous but is related to a 21

viscous resistance (Barry, 1974). After a sudden imposition of stress, the spring will eventually reach the strain, but the dashpot will retard the growth of the strain (Barnes et al., 1989). The strain in the spring is equal to that in the dashpot, and the total stress is the sum of the stress in the spring and in the dashpot (Barry, 1974). A removal of the stress allows the material to fully recover. When the stress is removed, the recovery of the strain is retarded. When depicting the viscoelastic behaviour of commercial pharmaceutical creams, single generalized Maxwell and Voigt models usually cannot represent the internal structure of the cream. The more complicated the structure of the cream, the more elements are needed to formulate a mechanical model. By connecting enough single elements of Maxwell and Voigt, even very complicated viscoelastic structures can be represented.

2.4.4 Viscoelastic test methods

Viscoelastic test methods should be used to get comprehensive information about the structural properties of pharmaceutical creams. With viscoelastic rheological measurements the elastic properties of solids and the viscous properties of liquids, and the ratios between these two, can be determined. Understanding of the internal structure of the cream, and its changes under the stress, enables one to predict the stability and usability of the cream. Two different types of methods are available to determine the linear viscoelastic behaviour of a material: dynamic and static methods (Barnes et al., 1989). The dynamic methods involve the application of harmonically varying stress or strain. The static methods involve the imposition of a step change in the stress or strain and the observation of the subsequent development of the strain or stress as a function of time. Analyses of the viscoelastic materials are designed not to destroy the structures, so that the measurements can provide information about the intermolecular and interparticle forces of the materials (Martin, 1993).

Dynamic methods

In oscillation tests the viscoelastic samples are subjected to sinusoidal oscillation stresses or oscillation strains (Schramm, 1994). In an oscillation test, a sinusoidally varying stress/strain of a fixed or varying period is applied to the sample, and the resulting sinusoidal strain/stress is measured. For a linearly viscoelastic material, the amplitude of stress is proportional to the 22

amplitude of strain, and the stress alternates sinusoidally at the same frequency, but it is out of phase with the strain (Fig. 12) (Barry, 1974). The in-phase response, at phase angle 0, shows that the sample is ideal elastic material, and the out-of-phase response, at phase angle 90, shows that the sample is ideal viscous material (Davis, 1974). With viscoelastic materials the phase angle, , between the applied stress/strain and the resulting strain/stress is between 0 and 90 (Schramm, 1994). The phase angle is a good indicator of the overall viscoelastic nature of the material (Tamburic et al., 1996).

Figure 12. Oscillation curves of a viscoelastic fluid, a Newtonian liquid and an ideal solid (Schramm, 1994). The stress is obtained as a function of strain. The behaviour is presented with a four-element viscoelastic model. is the angular deflection in radians; corresponds to 180, 2 to a full circle of 360.

When the amplitude ratio and the phase angle are measured, the parameters representing viscoelastic behaviour can be calculated. A complex shear modulus, G*, represents the total resistance of a material against the applied stress (Schramm, 1994). The complex modulus G* is separated into the in-phase and out-of-phase components

G* = G + iG,

[6]

23

where the storage modulus G is the in-phase (real) component and the loss modulus G is the out-of-phase (imaginary) component (Barry, 1974). The storage modulus G gives information about the elastic properties of the material

G = ( / ) cos

[7]

where is the stress, is the strain and the phase angle (Marriott, 1988). The greater the G, the more elastic are the characteristics of the material. The loss modulus G gives information about the viscosity of the material

G = ( / ) sin .

[8]

The greater the G, the more viscous are the characteristics of the material. The storage modulus represents the spring-like deformation in mechanical models, while the loss modulus represents the dashpot-like deformation. A loss tangent, tan , is obtained by dividing Equation 8 with Equation 7

tan = G / G.

[9]

The loss tangent is the ratio between viscous and elastic properties, showing which one is the dominant one. With a tan value of 1, the elastic and viscous properties of the material are equal. The smaller the loss tangent is, the more elastic is the material (Davis, 1971). Tan is often the most sensitive indicator of various molecular motions within the material (Radebaugh and Simonelli, 1983).

24

Static methods

Static methods are either creep tests at constant stress or relaxation tests at constant strain (Barnes et al., 1989). The creep test is used far more often than the relaxation test. In the creep test a constant stress is applied and the strain of a sample is determined as a function of time (Davis, 1969a). In the relaxation test the sample is subjected to a predetermined strain, and the stress required to maintain this strain is measured as a function of time (Marriott, 1988). During the creep test the strain is measured as a function of applied stress and presented in terms of compliance, J. The compliance is calculated from the measured strain, , and the applied stress, ,

J (t) = (t) / .

[10]

The smaller the J, the more elastic are the characteristics of the material. The creep curve of pharmaceutical semisolids can usually be split up into three separate regions: the instantaneous elastic region representing elastic stretching of primary structural bonds, the curved viscoelastic region representing the orientation of crystals or droplets due to the breaking and reforming of secondary bonds, and the viscous flow (Fig. 13) (Barry and Warburton, 1968; Davis, 1969a; Davis, 1974). In Fig. 13, the region A-B corresponds to the perfectly elastic movement as a result of the Maxwell spring (Schramm, 1994). The region BC corresponds to the gradual increase in the strain curve, viscoelastic behaviour, that is related to the Voigt elements. The region C-D corresponds to the response of an ideal Newtonian fluid related to the Maxwell dashpot. The instantaneous elastic region recovers totally and the viscoelastic region partly (Barry and Eccleston, 1973a). The viscous region does not recover (Davis, 1969a). In the recovery curve, in Fig. 13, the region D-E corresponds to the instantaneous elastic recovery, and it is equivalent to the region A-B. The region E-F represents the elastic recovery and it is equivalent to B-C. The unrecoverable part of the curve corresponds to the viscous flow. The unrecoverable viscous flow represents the irreversibly broken internal bonds in the material.

25

Figure 13. Creep recovery curve of wool fat as a function of time (Martin, 1993; reproduced from Davis, 1974). The structure is presented with a six-element viscoelastic model. Region A-B corresponds to instantaneous elastic response, region B-C to viscoelastic response and region C-D to viscous response. Regions D-E and E-F correspond to instantaneous elastic and elastic recoveries, respectively.

If the sample is tested under conditions within the linear viscoelastic region, the elements that cause an elastic response will give an equal contribution to both the creep and recovery phases (Schramm, 1994).

26

3. AIMS OF THE STUDY

The aim of the present study was to investigate the rheological (elastic, viscoelastic and viscous) properties of pharmaceutical creams containing sorbitan fatty acid ester surfactants with dynamic, static and time-dependent viscosity measuring methods. The aim was to gain understanding about the effects of the structural differences of single surfactants and the viscoelastic properties of pharmaceutical creams.

The specific aims of the present study were:

1.

To determine the effects of the structure, i.e. the length and the double bonded hydrocarbon chains of the sorbitan fatty acid ester surfactants, on the rheological properties of pharmaceutical creams,

2.

To determine the effects of the concentration of the surfactant and the volume of the inner phase on the rheological properties of pharmaceutical creams containing sorbitan fatty acid ester surfactants, and

3.

To determine the effects of the storage time and storage conditions on the rheological properties of pharmaceutical creams containing sorbitan fatty acid ester surfactants.

27

4. EXPERIMENTAL

4.1 Materials

The surfactants used were non-ionic sorbitan fatty acid esters (Fluka, Switzerland; Dr W. Kolb AG, Germany): saturated sorbitan monolaurate (III, IV), sorbitan monopalmitate (III, IV) and sorbitan monostearate (III, IV), and unsaturated sorbitan monooleate (I - IV) and sorbitan trioleate (I, II) (Fig. 14). The saturated sorbitan monoesters differ from each other by the length of the hydrocarbon chain. The unsaturated sorbitan monooleate differs from the saturated sorbitan monostearate, as it has a double bond in the hydrocarbon chain. Instead of the one double bonded hydrocarbon chain of sorbitan monooleate, sorbitan trioleate contains three double bonded hydrocarbon chains.
a)
b)

OH

OH

OH

R4

O OH
O R1 = O C O R2 = O C O R3 = O C O R4 = O C

O R4

R4

Figure 14. (a) Structures of sorbitan monoesters: R1 = sorbitan monolaurate, R2 = sorbitan monopalmitate, R3 = sorbitan monostearate and R4 = sorbitan monooleate. (b) Sorbitan trioleate contains three R4 hydrocarbon chains.

28

The co-surfactants used (I, II) were non-ionic polyethylene glycol soya sterols (Henkel KGaA, Germany): polyethylene glycol 10 soya sterol (PEG 10 soya sterol) and polyethylene glycol 25 soya sterol (PEG 25 soya sterol). The cream formulations contained either sorbitan fatty acid ester surfactants alone (III, IV) or both sorbitan fatty acid ester surfactant and co-surfactant (I, II). In the three-component creams (III, IV) the oil phase was isopropyl myristate (Henkel KGaA, Germany). In the creams containing the co-surfactants (I, II), the other components were caprylic/capric triglyceride (Hls AG, Germany), isopropyl palmitate (Henkel KGaA, Germany), glycerine (85%) (Henkel KGaA, Germany), cetostearyl alcohol (Henkel KGaA, Germany), methylparaben (Nipa Laboratories, Great Britain) and propylparaben (Nipa Laboratories, Great Britain). The materials were used as received without further purification. Each cream contained purified water as the aqueous phase. Study creams were formulated based on the literature (I, II) or the comprehensive pre-tests (III, IV).

4.2 Methods

4.2.1 Cream preparation (I-IV)

The creams were prepared on a laboratory scale with batch sizes of 3 kg (I, II) or 100 g (III, IV). The preparation was performed using instrumented mixing pan (Molto-Mat Universal 5, O. Krieger, Muttenz, Swizerland) (I, II), u-blade blender (RW 16 basic, Ika Labortechnik, Janke & Kunkel, Staufen, Germany) with Ultra-Turrax (TP 18/10, Ika-Werk, Janke & Kunkel, Staufen, Germany) (III), or only Ultra-Turrax mixer (T 8, Ika Werke GmbH & Co. KG, Staufen, Germany) (IV). Phases with the same temperature (75C or 80C) were combined and several homogenization steps were performed during the cream preparation/cooling period. The preparation was continued until the cream reached the temperature of 20C or 30C. Within each study, the creams were prepared using the same manufacturing method.

29

4.2.2 Determination of rheological properties (I-IV)

Rheological measurements were performed with controlled stress rheometers: StressTech rheometer (Rheological Instruments AB, Lund, Sweden, Stress RheoLogic Basic software, version 2.2, Rheologica Instruments, Lund, Sweden) (I III) and Bohlin CS rheometer (Bohlin Rheologi AB, Lund, Bohlin CS Rheometer software version 4.03) (IV). The used systems were parallel plate (I, II) or concentric cylinder (III, IV) systems. In all the measurements the temperature of the base plate was adjusted to 25C. The rheological measurements were performed at least in triplicate from separate cream samples. Dynamic oscillation stress sweep test was used to determine the linear viscoelastic region and the viscoelastic properties of the creams (IIV). In the oscillation stress sweep test, the cream was exposed to a growing applied stress; the stress was increased until the structure of the cream was broken down. During the test the frequency was kept constant (1 Hz). The three main parameters determined in the dynamic oscillation tests were storage modulus G, loss modulus G and loss tangent tan . The end point of the linear viscoelastic region was determined as a stress, when the G value was dropped 10% from the linear level. At the stress of crossing over point, the G was equal in value to the G and the value of tan was 1 (Gasperlin et al., 1997). Dynamic oscillation frequency sweep test was used to determine the capability of the creams to resist structural changes under the increased frequency (I, II). The frequency was increased from 0.01 Hz to 30 Hz and the stress was kept constant. The used stress was in the linear viscoelastic region of each cream. Static creep recovery test was used to determine the viscoelastic properties of the creams divided into elastic, viscoelastic and viscous properties (III, IV). The sample was exposed to the stress either 120 or 126 s and the strain recovery was registered either up to 224 or 360 s. The parameters in the static test were creep and recovery compliance J. The creep recovery tests were performed at a stress in the linear viscoelastic region of each cream. In the viscosity test, the resistance to flow was measured during both the increasing (up curve) and decreasing (down curve) phases of stress/rate (I, III, IV). The applied ranges of stress/rate were selected based on the consistencies of the creams.

30

4.2.3 Determination of interfacial tension (IV)

The interfacial tension between oil and water was measured using du Nouy tensiometer (KSV Sigma 70, KSV, Finland) with a platinum ring (IV). From the interfacial tension plots, the critical micelle concentration (cmc) was extrapolated at the intersection point of two linear portions of the interfacial tension vs. log concentration plots. The areas per molecules of surfactants, just below the cmc (Acmc), were calculated by means of the Gibbs adsorption isotherm (Equations [1] and [2]). Interfacial measurements were performed in triplicate from separate samples. Measurements were performed at a temperature of 22.0 0.5C.

4.2.4 Determination of droplet size distribution (I-III)

Droplet sizes were determined using a light microscope (Olympus BX40F, Olympus Optical Ltd., Tokyo, Japan). A 10% suspension with the material of continuous phase was prepared, and the measurement was performed immediately after sample preparation. The diameter was determined from 200 (I, II) or 500 (III) randomly chosen droplets. Droplet size determinations were performed in dublicate from separate samples.

4.2.5 Determination of conductivity (I, II)

Conductivity was measured with a portable conductivity instrument (Mettler Check Mate 90, Mettler Toledo, Essex, USA). Conductivity determinations were performed at least in triplicate from separate samples with an ion sensitive transistor.

31

5. RESULTS AND DISCUSSION

5.1 Effect of the length of the saturated hydrocarbon chain on the rheological properties of the creams (III, IV)

The effect of the saturated hydrocarbon chain length was studied with the corresponding three-component cream formulations, in which the only variable was the used sorbitan fatty acid ester surfactant; it was either sorbitan monolaurate, sorbitan monopalmitate or sorbitan monostearate. The studied creams were o/w creams and they contained either 0.042 mol (IV, Table 3) or 0.017 mol (III, Table 1) of the surfactant. Sorbitan monostearate with the longest hydrocarbon chain formed the most elastic structure for the creams. The most elastic structure can be seen from the greatest G values and the widest linear viscoelastic region of the creams in the oscillation stress sweep test (Fig. 15). Sorbitan monopalmitate with a two carbons shorter hydrocarbon chain formed clearly less elastic, but still linearly viscoelastically behaving creams (Fig. 15). The G values of the creams containing sorbitan monopalmitate were small and the linear viscoelastic region was clearly narrower (up to 6 Pa) than that of the creams containing sorbitan monostearate (up to 102 Pa). The creams containing sorbitan monolaurate, with a six and four carbons shorter hydrocarbon chain than the ones in sorbitan monostearate and sorbitan monopalmitate, respectively, were viscous without a linear viscoelastic region (Fig. 15). The less elastic behaviour of the creams containing sorbitan monopalmitate, as compared to the creams containing sorbitan monostearate, can also be seen from the greater tan value in the linear viscoelastic region (Table 1). Based on the values of the tan , both the sorbitan monopalmitate and sorbitan monostearate containing creams could be classified as elastic (tan < 1). The smaller the tan , the more rubbery or elastomeric is the behaviour (Ferry, 1980).

32

Storage modulus G' (Pa)

Sorbitan monolaurate

4 3 2 1 0 0 20 40 Shear stress (Pa)

12000 Storage modulus G(Pa) 10000 8000 6000 4000 2000 0 0

Sorbitan monopalmitate Sorbitan monostearate

25

50

75

100

125

150

Shear stress (Pa) Figure 15. Storage modulus (G) values of the creams containing 0.042 mol of sorbitan fatty acid ester surfactant (IV, Table 3) in the oscillation stress sweep test. Standard deviations (sd) of the G values in the linear viscoelastic region were 1100 Pa for the creams containing sorbitan monostearate and 700 Pa for the creams containing sorbitan monopalmitate. Sorbitan monolaurate containing creams did not have a linear viscoelastic region. Note the different scaling of axes. (n=3).

Table 1. Storage modulus (G) and loss tangent (tan ) values in the linear viscoelastic regions, and the end point values of the linear viscoelastic regions of the creams containing 0.017 mol of sorbitan fatty acid ester surfactant (III, Table 1) in the oscillation stress sweep test. (n=3).

Parameter

Creams containing sorbitan monopalmitate 224 35 0.454 0.020

Creams containing sorbitan monostearate 2050 190 0.232 0.003

G (Pa) Tan End point of the linear viscoelastic region (Pa)

0.7 0.2

33.9 15.1

33

The results of the creep recovery tests supported the results of the oscillation stress sweep tests. The compliances, J, of the creams containing sorbitan monostearate were very low, clearly lower than the compliances of the creams containing sorbitan monopalmitate (Fig. 16).

Sorbitan monopalmitate

8 7 6
J (10 -4 1/Pa)

10 8 6 4 2 0 0 50 100 150 200 250 Time (s)

5 4 3 2 1 0 0 50
Time (s)
Sorbitan monostearate

100

J (10 1/Pa)

-3

150

200

250

stress applied

stress removed

Figure 16. Compliance (J) values of the creams containing 0.042 mol of sorbitan fatty acid ester surfactant (IV, Table 3) in the creep recovery test. Standard deviations (sd) were < 0.0003 1/Pa for the creams containing sorbitan monostearate and < 0.003 1/Pa for the creams containing sorbitan monopalmitate. The applied stresses were in the linear viscoelastic regions of the creams. Note the different scaling of the axes. (n=3).

It has been concluded that if the sample is tested in the linear viscoelastic region, the elements that cause an elastic response will give an equal contribution to both the creep and the recovery phases (Schramm, 1994). However, in the creams containing sorbitan monostearate, the instantaneous recovery was almost twice as large as the instantaneous strain at the beginning of the creep phase (Fig. 16). This occurred despite the fact that the stress (40 Pa) used in the creep recovery test was clearly in the linear viscoelastic region (up to 102 Pa) of the creams. After the instantaneous recovery, neither the creams containing sorbitan 34

monostearate nor those containing sorbitan monopalmitate showed practically any additional recovery (Fig. 16). In the viscosity test, the creams containing sorbitan monostearate resisted the applied stress the most (Fig. 17). The yield value of the creams was about 75 Pa. The most thixotropic creams were the creams containing sorbitan monopalmitate with a pseudoplastic behaviour in the up curve (Fig. 17). The very beginning of the down curve corresponded with dilatant behaviour, and the rest of the down curve with pseudoplastic behaviour (Fig. 17). The up curve of the creams containing sorbitan monolaurate exhibited Newtonian flow (Fig. 17). This indicates that the droplets did not aggregate to any significant extent due to the impact of the applied stress (Sherman, 1968).

120 100 Shear rate (1/s) 80

Sorbitan monopalmitate Sorbitan monostearate

Sorbitan monolaurate Shear rate (1/s) 500 400 300 200 100 0 0 1 2 3 Shear stress (Pa)

60 40 20 0 0 50 100 Shear stress (Pa)

150

200

Figure 17. Viscosity curves of the creams containing 0.042 mol of sorbitan fatty acid ester surfactant (IV, Table 3). Standard deviations (sd) at the maximum shear stresses were 1 1/s for the creams containing sorbitan monolaurate, 40 1/s for the creams containing sorbitan monopalmitate, and 1 1/s for the creams containing sorbitan monostearate. The up and down curves of the creams containing sorbitan monopalmitate and sorbitan monolaurate are marked with arrows. Note the different scaling of the axes. (n=3).

The increase in the length of the saturated hydrocarbon chain increased the elastic nature of the creams. Corresponding findings have been observed in earlier studies as well, indicating that the increasing chain length of the surfactant, up to a certain limit, produced more 35

consistent creams (Barry and Eccleston, 1973a,b; Barry and Saunders, 1971; Eccleston and Beattie, 1988). In those studies the increases in consistency were concluded to be due to several reasons; due to the increasing number of hydrophobic alcohol-surfactant bonds, due to the formation of molecular complexes and smectic structures, and due to the decrease in the amount of unbound water in the system. In the present study, the study creams were simple three-component formulations. Based on this, the viscoelastic properties of the study creams were concluded to be mainly due to the interfacial properties of the surfactants. From the surfactants studied, sorbitan monostearate has smaller cmc and Acmc values than the molecules of sorbitan monopalmitate (IV, Table 2). Small values of cmc and Acmc enhance the condensed interfacial packing of the molecules. Sorbitan monostearate and sorbitan monopalmitate lowered the interfacial tension more effectively than sorbitan monolaurate (IV, Table 2). The conclusion that sorbitan monostearate and sorbitan monopalmitate have more condensed packing than sorbitan monolaurate is supported by the finding that the former two, with longer hydrocarbon chains, can better tolerate compression and are more readily associated with each other than sorbitan monolaurate (Peltonen and Yliruusi, 2000). Contrary to the vertical orientation of the surfactants with a long hydrocarbon chain, surfactant molecules with polar groups in the hydrocarbon chain prefer a more horizontal orientation at the interfaces (Rakshit et al., 1981). Due to the affinity of the short hydrocarbon chained sorbitan monolaurate for the aqueous phase (Opawale and Burgess, 1998b), the surfactant film of sorbitan monolaurate tends to be in an expanded horizontal state (Adamson, 1960). In the horizontal orientation the molecules require more space at interfaces, leading to looser packing and weaker chain-chain interactions between the molecules at the interfaces. In the present study, the increase in the length of the hydrocarbon chain of the saturated sorbitan fatty acid ester surfactant increased the elastic nature of the creams. Based on the minimal variables of the three-component study creams, the differences detected in the elastic nature of the creams were concluded to be mainly due to the differences in the length of the saturated hydrocarbon chain of the surfactants. From the results of the present study, it can be observed that the elastic nature of the creams increased as the cmc and Acmc values of the surfactants decreased.

36

5.2 Effect of the double bonds in the hydrocarbon chains on the rheological properties of the creams (I, IV)

The effect of the single double bonded hydrocarbon chain was studied with the corresponding three-component cream formulations, in which the only variable was the used surfactant; sorbitan monooleate and its saturated counterpart sorbitan monostearate. The studied creams were o/w creams and they contained 0.042 mol of the surfactant (IV, Table 3). In the oscillation stress sweep test, the creams containing sorbitan monostearate had wide linear viscoelastic regions with great values of G (Fig. 15). The creams containing sorbitan monooleate did not have elastic properties; the values of G were zero right from the beginning of stress application, as they were also in the creams containing sorbitan monolaurate in Fig. 15. In the viscosity test, the effect of the double bond could also be seen clearly. While the creams containing sorbitan monostearate showed high resistance against the applied stress (yield value of about 75 Pa), the creams containing sorbitan monooleate flowed almost like Newtonian liquids (Fig. 18).
Shear rate (1/s)

1,0

200 150 100 50 0 0 1

Sorbitan monooleate

Shear rate (1/s)

S hear s tres s (Pa)

0,5

Sorbitan monostearate 0,0 0 50 100 Shear stress (Pa) 150 200

Figure 18. Viscosity curves of the creams containing 0.042 mol of sorbitan fatty acid ester surfactant (IV, Table 3). Standard deviations (sd) at maximum shear stresses were 1 1/s for the creams containing sorbitan monostearate and 2 1/s for the creams containing sorbitan monooleate. The up and down curves are marked with arrows. Note the different scaling of the axes. (n=3). 37

The double bond in the hydrocarbon chain clearly decreased the elastic nature of the creams. As it was concluded in the previous section, the viscoelastic properties of simple threecomponent creams are caused mainly by the interfacial properties of the surfactants. Sorbitan monooleate lowered the interfacial tension less than sorbitan monostearate, and the double bond in the hydrocarbon chain almost doubled the molecular area, Acmc, of sorbitan monooleate, as compared to sorbitan monostearate (IV, Table 2). It has been concluded that the double bonded structure loosens the packing of the surfactants at the interfaces (Carlotti et al., 1995), and decreases the hydrophobic chain-chain interactions between the adjacent surfactant molecules and oil molecules (Feher et al., 1977). The effect of the number of double bonded hydrocarbon chains was studied with corresponding cream formulations: sorbitan monooleate or sorbitan trioleate (7% w/w) combined with co-surfactants PEG 10 soya sterol or PEG 25 soya sterol (12% w/w) (I, Table 1). Sorbitan trioleate had three double bonded hydrocarbon chains and formed more elastic structures for the creams than sorbitan monooleate that contains only one double bonded hydrocarbon chain (Table 2). The results of the oscillation stress sweep test showed that neither of the sorbitan fatty acid ester surfactants could form linearly viscoelastically behaving creams with the co-surfactant PEG 25 soya sterol.

Table 2. Storage modulus (G), loss tangent (tan ) and crossing over point values of creams containing 7% (w/w) of sorbitan fatty acid ester surfactant and 12% (w/w) of soya sterol surfactant (I, Table 1) in the oscillation stress sweep test. (n=6).

Parameter

Creams containing sorbitan monooleate + PEG 10 soya sterol 320 13 0.07 36 1.4

Creams containing sorbitan trioleate + PEG 10 soya sterol 960 32 0.07 104 6.1

Creams containing sorbitan monooleate + PEG 25 soya sterol 56 16 0.59 2.8 0.7

Creams containing sorbitan trioleate + PEG 25 soya sterol 120 19 0.56 8.1 1.9

G (Pa) Tan Crossing over point (Pa)

In the oscillation frequency sweep test, the G values of the creams containing sorbitan trioleate increased steadily implying structural changes under the growing frequency (Fig. 19). The decreasing tan values of the creams imply that the proportion of elastic properties increased as a function of frequency (Fig. 19). Which property is dominating, depends on the stress and the duration of stress application (Barnes et al., 1989). As the frequency is 38

increased, the time for straining is decreased and the elastic deformations become dominant; there is no time for the viscous deformations. Up to 10 Hz both creams could be classified as elastic, as the values of tan were < 1 (Fig. 19). When the values of G and tan stay stable under an increasing frequency, the cream resists the structural changes, which may imply good resistance against flocculation and coalescence and, hence, good stability during storage.

0,3

G(Pa)

0,1

10

100

Tan

Frequency (Hz)

Figure 19. Storage modulus (G) and loss tangent (tan ) values of the creams containing 7% (w/w) of sorbitan fatty acid ester surfactant and 12% (w/w) of PEG 10 soya sterol cosurfactant (I, Table 1) in the oscillation frequency sweep test. Standard deviations (sd) of the values of G and tan were < 38 Pa and < 0.03, respectively, for the cream containing sorbitan monooleate, and < 51 Pa and < 0.02, respectively, for the cream containing sorbitan trioleate. (n=6).

The creams containing sorbitan trioleate had smaller inner phase droplets than the creams containing sorbitan monooleate (I, Table 4). This supports the finding that the consistency is increased, when the particle size of the dispersed phase is decreased (Rieger, 1986). As the number of droplets is increased, the interactions between the droplets and the consistency are also increased. However, the droplet size distribution is not the only explanatory factor when considering the rheological properties of pharmaceutical creams. As commercial pharmaceutical creams are considered complex polydispersed multiple-phase systems containing additional phases of oil and water (Eccleston, 1990), the viscoelastic networks present in the continuous phases are mainly responsible for the consistency of these creams (Barry and Eccleston, 1973a).

39

1400 1200 1000 800 600 400 200 0 0,01

Sorbitan monooleate Sorbitan trioleate

0,2

0,1

0,0 0,01

0,1

1
Frequency (Hz)

10

100

It has been concluded that in the case of non-ionic surfactants, the stabilization of emulsions occurs through steric hindrance and hydrogen bonding (Fox, 1986). Boyd et al. (1972) reported that sorbitan trioleate does not form as condensely associated structures with polyoxyethylene sorbitan monopalmitate or polyoxyethylene sorbitan monolaurate as sorbitan monooleate. This was due to the more open configuration of the sorbitan trioleate at the interfaces. However, in the present study the increase in the number of double bonded hydrocarbon chains from one to three per molecule increased the elastic nature of the creams. The structural interactions between the single surfactants and their complementary hydrophile-lipophile balance (HLB), for example, have an influence on the effectiveness of the mixture of the surfactants. Shidona and Yoneyama (1980) showed that in the mixture of sorbitan fatty acid ester and polyoxyethylene sorbitan fatty acid ester type surfactants the creams were more stable when the differences in the HLB values of the mixed surfactants were small. In the present study sorbitan trioleate with a HLB value of 1.8 interacted with the hydrophilic co-surfactant PEG 10 soya sterol (a HLB value of 12) more effectively than sorbitan monooleate with a HLB value of 4.3. The structural interactions between the single surfactants and their complementary hydrophile-lipophile balances, for example, influence the effectiveness of the mixture of the surfactants. Contributing factors are also the orientation of the surfactant molecules at the interfaces and the compatibility of the surfactant mixture with the whole cream formulation. As concluded earlier, the droplet size distribution and the viscoelastic networks at the continuous phase also affect the rheological properties of the creams.

5.3 Effect of the concentration of the surfactant on the rheological properties of the creams (III)

The effect of the surfactant concentration was studied with the corresponding threecomponent cream formulations that contained 0.007 mol or 0.011 mol of sorbitan monolaurate, or 0.011 mol or 0.017 mol of sorbitan monostearate (III, Table 1). Sorbitan monolaurate containing creams were w/o creams with high internal phase volume, and sorbitan monostearate containing creams were o/w creams with low internal phase volume. In the viscoelastic measurements the increasing concentration of the surfactant increased the elastic nature of the creams containing either sorbitan monolaurate or sorbitan monostearate (Table 3). Although the sorbitan monostearate containing creams with the greater amount (0.017 mol) of the surfactant had greater G values, these creams had slightly greater tan values than the creams containing the smaller amount (0.011 mol) of the surfactant (Table 3). This indicates that the proportion of G was slightly more dominating in the creams containing the greater amount of sorbitan monostearate. With both the surfactants, similarly to 40

G values, also the end points of the linear viscoelastic regions and the crossing over points indicated greater elasticity in the creams that contained more of the surfactant (Table 3). An increase in the stress at the end point of the linear viscoelastic region and in the stress at the crossing over point is a sign of pronounced elastic properties.

Table 3. Values of the storage modulus (G), loss modulus (G), loss tangent (tan ), end points of the linear viscoelastic regions and crossing over points of the creams containing sorbitan fatty acid ester surfactant (III, Table 1) in the oscillation stress sweep test. (n=6).

Parameter

Creams containing sorbitan monolaurate 0.007 mol 212 10 3.5 0.1 0.016 0.001

Creams containing sorbitan monolaurate 0.011 mol 273 6 3.8 0.2 0.014 0.001

Creams containing sorbitan monostearate 0.011 mol 97 19 20 4.3 0.206 0.006

Creams containing sorbitan monostearate 0.017 mol 2050 192 475 42 0.232 0.003

G (Pa) G (Pa) Tan End point of the linear viscoelastic region (Pa) Crossing over point (Pa)

14.0 2.4 20.0 2.8

26 7 31 6

0.7 0.1 2.7 0.7

34 15 54 25

The results of the creep recovery tests supported the results of the oscillation stress sweep tests. In the creams containing sorbitan monolaurate the effect of the higher concentration of the surfactant and the increase of compliance were small during the stress application (Fig. 20). The recovery was complete after the removal of stress. In the creams containing sorbitan monostearate an increase in the concentration of the surfactant formed clear differences in the compliance values (Fig. 20).

41

6 5 J (10-2 1/Pa) 4 3 2 1 0 0 100 200 Time (s)

Sorbitan monolaurate 0.007 mol Sorbitan monolaurate 0.011 mol Sorbitan monostearate 0.011 mol Sorbitan monostearate 0.017 mol

300

400

stress applied

stress removed

Figure 20. Compliance (J) values of the creams containing sorbitan fatty acid ester surfactant (III, Table 1) in the creep recovery test. Standard deviations (sd) were < 0.0006 1/Pa for the creams containing sorbitan monolaurate, < 0.01 1/Pa for the creams containing 0.011 mol of sorbitan monostearate and < 0.001 1/Pa for the creams containing 0.017 mol of sorbitan monostearate. The applied stresses were in the linear viscoelastic regions of the creams. (n=6).

In the viscosity test the creams with the high internal phase volume, containing sorbitan monolaurate, behaved anti-thixotropically (Fig. 21). The up curve of the creams containing 0.007 mol of sorbitan monolaurate showed dilatant behaviour, indicating that the closely packed droplets started to take a more open form of packing during the shearing process. With sorbitan monolaurate containing creams the results of the viscosity tests did not support the results of the viscoelastic tests. The creams containing a smaller amount (0.007 mol) of surfactant resisted the shear rate more than the creams containing a greater amount (0.011 mol) of surfactant (Fig. 21). Only at the very beginning of the viscosity test the behaviours were opposite. The different behaviours can result from the different principles of the viscosity and viscoelasticity measurements; while the viscosity measurements brake down the internal structure of the cream, the viscoelastic measurements are designed not to destroy the structure but to provide information about the intermolecular forces of the cream. The creams containing sorbitan monostearate were pseudoplastic thixotropic systems (Fig. 21). The greatest yield value, about 100 Pa, was found in the creams containing a greater amount (0.017 mol) of sorbitan monostearate (Fig. 21). This supports the great G, end point of the

42

linear viscoelastic region and crossing over point values of these creams in the oscillation stress sweep test (Table 3).

600 Shear rate (1/s) 500 400 300 200 100 0 0 50 100 150 200 Shear stress (Pa) 250 300 350
Sorbitan monolaurate 0.007 mol Sorbitan monolaurate 0.011 mol Sorbitan monostearate 0.011 mol Sorbitan monostearate 0.017 mol

Figure 21. Viscosity curves of the creams containing sorbitan fatty acid ester surfactant (III, Table 1). Standard deviations (sd) were < 14.5 Pa and < 4.5 Pa for the creams containing 0.007 mol and 0.011 mol of sorbitan monolaurate and < 1.4 Pa and < 18.7 Pa for the creams containing 0.011 mol and 0.017 mol of sorbitan monostearate, respectively. The up curves are marked with arrows. Measurements were performed as a function of shear rate. (n=6).

In the present study, viscoelasticity measurements showed that increasing the concentration of the surfactant increased the elastic nature of the creams. An increased consistency due to the increasing concentration of the surfactant, up to a certain limit, has been noticed also in other studies (Barry, 1968; Barry and Saunders, 1970,1971,1972; Barry and Eccleston, 1973a,c; Eros and Urgi-Hunyadvari, 1979; Virtanen et al., 1993; Vasiljevic et al., 1994; Chanana and Sheth, 1995; Frster and Herrington, 1997). Generally, the greater the amount of the surfactant, the greater the possibility to form smaller inner phase droplets and a more complex structure to the continuous phase. In the present study there was a correlation between the droplet size distribution and the tan values of the creams with both the surfactants. The creams that had the smallest tan values, i.e. the creams containing 0.011 mol of sorbitan monolaurate or sorbitan monostearate, had the smallest droplets (III, Fig. 4).

43

5.4 Effect of the volume of the inner phase on the rheological properties of the creams (III)

The effect of the inner phase volume was studied with corresponding three-component cream formulations containing 0.011 mol of sorbitan monolaurate or sorbitan monostearate as the surfactant (III, Table 1). Sorbitan monolaurate containing w/o creams, with 87% (w/w) inner phase volume, formed a clearly more elastic structure for the creams than sorbitan monostearate containing o/w creams, with 9% (w/w) inner phase volume (Table 4). The creams containing sorbitan monolaurate could be classified as high internal phase ratio emulsion (HIPRE) creams, in which the volume of the inner phase exceeds the volume fraction of the maximum packing. In this case, the droplets just touch each other already in a rest situation (Pal, 1999a). Due to the high volume of the inner phase of the creams, sorbitan monolaurate could produce more elastic and more complex structure for the creams, despite its six carbons shorter hydrocarbon chain, than sorbitan monostearate. In addition to the greater inner phase volume of the sorbitan monolaurate containing w/o creams, oil as the continuos phase could also have an effect on the consistency difference between the creams. Both the creams could be classified as elastic (tan < 1) (Table 4).

Table 4. Storage modulus (G), loss modulus (G), loss tangent (tan ), end point of the linear viscoelastic region and crossing over point values of the creams containing 0.011 mol of sorbitan fatty acid ester surfactant (III, Table 1) in the oscillation stress sweep test. (n=6).

Parameter

Creams containing sorbitan monolaurate 273.2 5.5 3.8 0.2 0.014 0.001

Creams containing sorbitan monostearate 96.8 19.4 19.9 4.3 0.206 0.006

G (Pa) G (Pa) Tan End point of the linear viscoelastic region (Pa) Crossing over point (Pa) Volume of inner phase (%) Type of the cream

25.9 6.8 30.7 6.0 87 w/o

0.7 0.1 2.7 0.7 9 o/w

44

In the creep recovery test the creams containing sorbitan monolaurate behaved like a Hookean solid; after the instantaneous elastic response, there were no curved viscoelastic or linear viscous responses, and the creams recovered totally after the removal of stress (Fig. 22). The creams containing sorbitan monostearate had clear elastic, viscoelastic and viscous regions in their curve. Despite the great differences in the creep recovery curves, both creams containing 0.011 mol of the surfactant could be represented by the Burger viscoelastic model containing two springs and two dashpots (III, Fig. 5 and Table 4).
Sorbitan monolaurate Sorbitan monostearate

6 5 J (10 1/Pa) 4 3 2 1 0 0 50 100 150 200 Time (s) 250

-4

300

350

400

stress applied

stress removed

Figure 22. Compliance (J) values of the creams containing 0.011 mol of sorbitan fatty acid ester surfactant (III, Table 1) in the creep recovery test. Standard deviations (sd) were < 0.6x10-3 1/Pa for the creams containing sorbitan monolaurate and < 1.4x10-2 1/Pa for the creams containing sorbitan monostearate. The applied stresses were in the linear viscoelastic regions of the creams. (n=6).

In the viscosity test the creams containing sorbitan monolaurate had a clear yield value and they behaved anti-thixotropically (Fig. 23). The great amount of droplets in HIPRE creams caused shear-thickening effect for the creams. The creams containing sorbitan monostearate were pseudoplastic thixotropic systems.

45

600 Shear rate (1/s) 500 400 300 200 100 0 0 50 100 150 Shear stress (Pa)

Sorbitan monolaurate Sorbitan monostearate

200

250

Figure 23. Viscosity curves of the creams containing 0.011 mol of sorbitan fatty acid ester surfactant (III, Table 1). Standard deviations (sd) were < 4.5 Pa for the creams containing sorbitan monolaurate and < 1.4 Pa for the creams containing sorbitan monostearate. Up and down curves are marked with arrows. Measurements were performed as a function of shear rate. (n= 6-9).

Droplet size is known to be a very important variable in emulsion rheology (Barnes, 1994). In the present study, there was a correlation between droplet size distribution and viscoelastic properties. Creams with the smallest droplets, i.e. the creams containing sorbitan monolaurate (III, Fig. 4), behaved elastically (tan = 0.014). The creams containing sorbitan monostearate had larger droplets on average (III, Fig. 4), and the creams did not behave as elastically (tan = 0.206). It has been concluded that the volume of the inner phase affects the rheological properties of creams (Pal et al., 1986; Dahms, 1991; Jager-Lezer et al.; 1998, Pal, 1999a,b). Similarly to those studies, also in the present study the increase in the inner phase volume increased the elastic nature of the creams. In the present study, the increase in consistency was detected both in the viscoelasticity measurements and in the viscosity measurements.

5.5 Effect of storaging on the rheological properties of the creams (I, II)

The short-term stabilities of the creams were studied at room temperature (25C, RH 60%) at 1-2 days (0-sample), 14 and 28 days, and in cold (5C) and accelerated (40C, RH 75%) conditions at 28 days. The effect of storage was studied with corresponding cream 46

formulations of sorbitan monooleate or sorbitan trioleate (7% w/w) combined with cosurfactants PEG 10 soya sterol or PEG 25 soya sterol (12% w/w) (I, Table 1; II, Table 1). Sorbitan trioleate resulted in more elastic creams than sorbitan monooleate (Table 5). During the 28-day storage period at room temperature, the most consistent creams of each formulation were found at the beginning of the stability test (Table 5). A drop in the G value was the greatest in the creams containing sorbitan trioleate and PEG 10 soya sterol (Table 5). However, despite the greatest loss in elasticity, they were still clearly the most elastic creams after the 28-day storage period at 25C RH 60%.

Table 5. Values of storage modulus (G) as a function of storage time of the creams containing 7% (w/w) of sorbitan fatty acid ester surfactant and 12% (w/w) of soya sterol surfactant (I, Table 1) in the oscillation stress sweep test. Creams were stored at a room temperature (25C RH 60%). (n= 6-9).

Storage time

Creams containing sorbitan monooleate + PEG 10 soya sterol 319 12 274 25 224 12

Creams containing sorbitan trioleate + PEG 10 soya sterol 947 28 636 24 596 27

Creams containing sorbitan monooleate + PEG 25 soya sterol 60 16 43 3.3 42 5.2

Creams containing sorbitan trioleate + PEG 25 soya sterol 113 17 91 21 96 34

0-sample 14 days 28 days

Storing the creams for 28 days in three different storage conditions attenuated the differences in the elasticities. Sorbitan trioleate formed slightly more elastic creams than sorbitan monooleate when the highest values of G were compared irrespective of the storage conditions (Fig. 24).

47

800

5C 25C RH 60% 40C RH 75%

600 G (Pa)

400

200

0 Sorbitan monooleate + Sorbitan trioleate + Sorbitan monooleate + Sorbitan trioleate + PEG 10 soya sterol PEG 10 soya sterol PEG 25 soya sterol PEG 25 soya sterol Formulation

Figure 24. Values of storage modulus (G) of the creams containing 7% (w/w) of sorbitan fatty acid ester surfactant and 12% (w/w) of soya sterol surfactant (II, Table 1) after 28 days of storage in three different storage conditions. The error bars represent standard deviations ( sd) for six oscillation stress sweep tests.

In the oscillation frequency sweep test the most viscoelastic creams, sorbitan monooleate or sorbitan trioleate combined with PEG 10 soya sterol, behaved similarly after the 28 days of storage (Fig. 25). The most viscoelastic sorbitan monooleate and sorbitan trioleate containing creams were stored at 5C and 25C RH 60%, respectively. With both creams the G values increased with the increasing frequency (Fig. 25). Despite that, the values of tan decreased up to 1 Hz, after which the proportion of the G started to increase. Up to 10 Hz both creams could be classified as elastic (tan < 1).

48

Sorbitan monooleate Sorbitan trioleate 0.20 Tan (Pa)

0.15 0.10 0.05 0.00 0.01 0.1 1 10

Frequency (Hz)

Figure 25. Values of the storage modulus (G) and loss tangent (tan ) of the most viscoelastic creams containing 7% (w/w) of sorbitan fatty acid ester surfactant and 12% (w/w) of PEG 10 soya sterol surfactant (II, Table 1) after 28 days of storage in the oscillation frequency sweep test. Sorbitan monooleate containing creams were stored at 5C and the sorbitan trioleate containing creams at 25C RH 60%. The error bars present standard deviations (sd) for each value. The error bars of tan values are within the point marks. (n=6).

It has been generally concluded that the viscoelastic nature of a cream is a good indicator of its stability (Frster and Herrington, 1997; Zografi, 1982; Gasperlin et al., 1998). The temperature-dependent breakdown is the most common reason for phase separation. Both the decrease of temperature (Kallioinen et al., 1994) and the increase of temperature (Davis, 1969b; Santos Magalhaes et al., 1991) have been reported to enhance the phase separation of pharmaceutical creams. The decrease in consistency detected when storage temperature rises can be a result of a temperature-dependent breakdown in the structures that are responsible for the viscoelastic properties (Davis, 1969b). The increase in consistency detected when the storage temperature rises may be due to the increased Brownian movement in the system (Barry and Saunders, 1970). This is believed to form new linkages and to strengthen the cream network by bringing the free ends of strands into contact with the main meshwork (Barry and Saunders, 1970). In the present study the most viscoelastic creams, containing sorbitan trioleate and PEG 10 soya sterol as surfactants, had the smallest droplets (II, Table 3). In addition, in the present study the conductivity test mainly supported the observation that when the amount of structurally bound water is decreased, and thus conductivity increased, the consistency is decreased (I, Table 5; II, Table 4) (Eccleston and Beattie, 1988). It has been suggested that the later the conductivity starts to rise and the smaller the rise is, the more stable the cream is going to be (Virtanen et al., 1993). In the present study, based on this observation, it could be 49

predicted that the most stable creams in long-term would be the creams containing sorbitan monooleate and PEG 10 soya sterol when stored at 25C RH 60% and at 5C and the creams containing sorbitan trioleate and PEG 10 soya sterol when stored at 25C RH 60% (II, Table 4). In these creams the conductivity values stayed stable through the short-term stability study.

50

6. CONCLUSIONS

The conclusions drawn from the present study were:

1.

An increase in the length of the saturated hydrocarbon chain of the surfactant (sorbitan monolaurate vs. sorbitan monopalmitate vs. sorbitan monostearate) increased the elastic nature of the creams, as did an increase in the number of double bonded hydrocarbon chains in the surfactant molecule (sorbitan monooleate vs. sorbitan trioleate). A double bond in the hydrocarbon chain of the surfactant (sorbitan monooleate vs. sorbitan monosterate) decreased the elastic nature of the creams.

2.

An increase in the concentration of the surfactant (0.007 mol vs. 0.011 mol and 0.011 mol vs. 0.017 mol) and an increase in the volume of the inner phase (9% vs. 87%) increased the elastic nature of the creams containing sorbitan fatty acid ester surfactants.

3.

Storage time (up to 28 days) decreased and storaging the creams at three different conditions (5C vs. 25C RH 60% vs. 40C RH 75%) levelled off the differences in the elastic properties of the creams containing sorbitan fatty acid ester surfactants.

Furthermore, it was concluded that rheological measurements gave comprehensive information about the structural (elastic, viscoelastic and viscous) properties of the pharmaceutical creams containing sorbitan fatty acid ester surfactants. Dynamic and static viscoelastic measurements and time-dependent viscosity measurements complement each other as physical testing methods of creams. Rheological measurements can be highly recommended for the quality evaluation of pharmaceutical creams.

51

REFERENCES

Adamson, A.W., 1960. Physical Chemistry of Surfaces. Interscience Publishers Inc., New York, pp. 124-138. Attwood, D., Florence, A.T., 1983. Surfactant Systems. Their Chemistry, Pharmacy and Biology. Chapman and Hall Ltd., London. Autian, J., 1966. Emulsions. In: Martin, E.W., Hoover, J.E. (Ed.), Husas Pharmaceutical Dispending. 6th Ed., Mack Publishing Company, Pennsylvania, pp. 223-258. Barnes, H.A., Hutton, J.F., Walters, K., 1989. An Introduction to Rheology. Elsevier, Amsterdam. Barnes, H.A., 1994. Rheology of emulsions - a review. Colloids Surfaces 91, 89-95. Barry, B.E., 1968. The self bodying action of the mixed emulsifier sodium dodecyl sulfate / cetyl achohol. J. Colloid Interface Sci. 28, 82-91. Barry, B.W., 1974. Rheology of Pharmaceutical and Cosmetic Semisolids. In: Bean, H.S., Beckett, A.H., Carless, J.E. (Ed.), Advances in Pharmaceutical Sciences. Vol 4, Academic Press Inc., London, pp. 1-72. Barry, B.W., 1983. Rheology of Dermatological Vehicles. In: Dermatological Formulations, Percutaneous Absorption. Marcel Dekker, Inc., New York, pp. 351-407. Barry, B.W., Eccleston, G.M., 1973a. Influence of gel networks in controlling consistency of o/w emulsions stabilised by mixed emulsifiers. J. Texture Stud. 4, 53-81. Barry, B.W., Eccleston, G.M., 1973b. Oscillatory testing of o/w emulsions containing mixed emulsifiers of the surfactant/long chain alcohol type: influence of surfactant chain length. J. Pharm. Pharmacol. 25, 394-400. Barry, B.W., Eccleston, G.M., 1973c. Oscillatory testing of o/w emulsions containing mixed emulsifiers of the surfactant-long chain alcohol type: self-bodying action. J. Pharm. Pharmacol. 25, 244-253. Barry, B.W., Saunders, G.M., 1970. Rheology of systems containing cetrimide - cetostearyl alcohol: variation with temperature. J. Pharm. Pharmacol. 22, 139-146.

52

Barry, B.W., Saunders, G.M., 1971. The self-bodying action of alkyltrimethylammonium bromides/cetostearyl alcohol mixed emulsifiers; influence of quaternary chain length. J. Colloid Interface Sci. 35, 689-705. Barry, B.W., Saunders, G.M., 1972. Rheology of systems containing cetomacrocol 1000 cetostearyl alcohol. I. Self-bodying action. J. Colloid Interface. Sci. 38, 616-625. Barry, B.W., Warburton, B., 1968. Some rheological aspects of cosmetics. J. Soc. Cosmet. Chemists 19, 725-744. Becher, P., 1957. Emulsions: Theory and Practice. Reinhold Publishing Corporation, New York. Billany, M.R., 1988. Emulsions. In: Pharmaceutics: The Science of Dosage Form Design. Aulton, M.E. (Ed.), Churchill Livingstone, New York. 282-299. Bjerregaard, S., Pedersen, H., Vedstesen, H., Vermehren, C., Sderberg, I., Frokjaer, S., 2001. Parenteral water/oil emulsions containing hydrophilic compounds with enhanced in vivo retention: formulation, rheological characterisation and study of in vivo fate using whole body gamma-scintigraphy. Int. J. Pharm. 215, 13-27. Block, L.H., 1989. Emulsions and Microemulsions. In: Lieberman, H.A., Rieger, M.M., Banker, G.S. (Ed.), Pharmaceutical Dosage Forms: Disperse Systems, Vol. 2, Marcel Dekker, Inc., New York, pp. 335-378. Boyd, J., Parkinson, C., Sherman, P., 1972. Factors affecting emulsion stability, and the HLB concept. J. Colloid Interface Sci. 41, 359-370. Briceno, M.I., 2000. Rheology of Suspensions and Emulsions. In: Nielloud, F., MartiMestres, G. (Ed.), Drugs and Pharmaceutical Sciences; Pharmaceutical Emulsions and Suspensions, Vol. 105, Marcel Dekker, Inc., New York, pp. 557-607. Carlotti, M.E., Pattarino, F., Gasco, M.R., Cavalli, R., 1995. Use of polymeric and nonpolymeric surfactants in o/w emulsion formulation. Int. J. Cosmet. Sci. 17, 13-25. Chanana, G.D., Sheth, B.B., 1995. Particle size reduction of emulsions by formulation designII: Effect of oil and surfactant concentration. PDA J. Pharm. Sci. & Techn. 49, 71-76. Cooper, J.W., Gunn, C., 1950. Tutorial Pharmacy. 4th Ed., Pitman Medical Publishing Company Ltd., London, pp. 241-268. Cosgrove, F.P., 1963. Emulsions. In: Burlage, H.M., Lee, C.O., Rising, L.W. (Ed.), Physical and Technical Pharmacy. The McGraw-Hill Book Company, Inc., New York, pp. 539-590. 53

Dahms, G., 1991. Viscosity regulation by surfactants. Int. J. Cosmetic Sci. 13, 43-50. Davis, S.S., 1969a. Viscoelastic properties of pharmaceutical semisolids, I: Ointment bases. J. Pharm. Sci. 58, 412-417. Davis, S.S., 1969b. Viscoelastic properties of pharmaceutical semisolids, II: Creams. J. Pharm. Sci. 58, 418-421. Davis, S.S., 1971. Viscoelastic properties of pharmaceutical semisolids III: Nondestructive oscillatory testing. J. Pharm. Sci. 60, 1351-1355. Davis, S.S., 1974. Is pharmaceutical rheology dead? Pharm. Acta Helv. 49, 161-168. Davis, H.T., 1994. Factors determining emulsion type: hydrophile-lipophile balance and beyond. Colloid. Surf. Physicochem. Eng. Aspects 91, 9-24. Dunker, M.F.W., 1960. Colloids, Emulsions and Suspensions. In: Sprowls, J.B. (Ed.), American Pharmacy, Textbook of Pharmaceutical Principles, Processes and Preparations. 5th Ed., J.B. Lippincott Company, Philadelphia, pp. 107-148. Eccleston, G.M., 1986a. Application of emulsion stability theories to mobile and semisolid o/w emulsions. Cosmet. Toiletries 101, 73-92. Eccleston, G.M., 1986b. The microstructure of semisolid creams. Pharm. Int., March 1986, 63-70. Eccleston, G.M., 1990. Multiple-phase oil-in-water emulsions. J. Soc. Cosmet. Chem. 41, 122. Eccleston, G.M., Beattie, L., 1988. Microstructural changes during the storage of systems containing cetostearyl alcohol/polyoxyethylene alkyl ether surfactants. Drug Dev. Ind. Pharm. 14, 2499-2518. Eros, I., Ugri-Hunyadvari, E., 1979. Investigation of the rheological characteristics of ointment gels containing emulsifiers and emulsion-type ointments. Cosmet. Toiletries 94, 6779. Feher, A.I., Collins, F.D., Healy, T.W., 1977. Mixed monolayers of simple saturated and unsaturated fatty acids. Austr. J. Chem. 30, 511-519. Ferry, J.D., 1980. Viscoelastic Properties of Polymers. 3rd Ed., John Wiley & Sons Inc., New York, pp. 33-55.

54

Fox, C., 1986. Rationale for the selection of emulsifying agents. Cosmet. Toiletries 101, 2544. Frster, A.H., Herrington, T.M., 1997. Rheology of siloxane-stabilized water in silicone emulsions. Int. J. Cosmet. Sci. 19, 173-191. Garrett, E.R., 1965. Stability of oil-in-water emulsions. J. Pharm. Sci., 54, 1557-1570. Gaspar, L.R., Maia Campos, P.M.B.G., 2003. Rheological behaviour and the SPF of sunscreens. Int. J. Pharm. 250, 35-44. Gasperlin, M., Kristl, J., Smid-Korbar, J., 1997. Viscoelastic behaviour of semi-solid w/o emulsion systems. STP Pharma Sci. 7 (2), 158-163. Gasperlin, M., Tusar, L., Tusar, M., Kristl, J., Smid-Korbar, J., 1998. Lipophilic semisolid emulsion systems: viscoelastic behaviour and prediction of physical stability by neural network modelling. Int. J. Pharm. 168, 243-254. Griffin,W.C., 1949. Classification of surface-active agents by HLB. J. Soc. Cosmet. Chem. 1, 311-326. Groves, M.J., de Galindez, D.A., 1976. Rheological characterization of self-emulsifying oil/surfactant systems. Acta Pharm. Suec. 13, 353-360. Jager-Lezer, N., Tranchant, J-F., Alard, V., Vu, C., Tchoreloff, P.C., Grossiord, J-L., 1998. Rheological analysis of highly concentrated w/o emulsions. Rheol. Acta 37, 129-138. Jiao, J., Burgess, D.J., 2003. Rheology and stability of water-in-oil-in-water multiple emulsions containing Span 83 and Tween 80. AAPS Pharm. Sci. 5, 62-73. Junginger, H., Akkermans A.A.M.D., Heering W., 1984a. The ratio of interlamellary fixed water to bulk water in o/w creams. J. Soc. Cosmet. Chem. 35, 45-57. Junginger, H.E., 1984b. Colloidal structures of o/w creams. Pharmaceutisch Weekblad Scientific Edition, 6, 141-149. Kallioinen, S., Helenius, K., Yliruusi, J., 1994. Physical characteristics of some emulsion creams in which the length of ethylene oxide chain of emulsifier varies. Pharmazie 49, 750756. Kanouni, M., Rosano, H.L., Naouli, N., 2002. Preparation of a stable double emulsion (W1/O/W2): role of the interfacial films on the stability of the system. Adv. Colloid Interface Sci. 99, 229-254. 55

Kawashima, Y., Hino, T., Takeuchi, H., Niwa, T., Horibe, K., 1991. Rheological study of o/w/o emulsion by a cone-and-plate viscometer: Negative thixotropy and shear-induced phase inversion. Int. J. Pharm. 72, 65-77. Kawashima, Y., Hino, T., Takeuchi, H., Niwa, T., 1992. Stabilization of water/oil/water multiple emulsion with hypertonic inner aqueous phase. Chem. Pharm. Bull., 40 (5), 12401246. Kayes, J.B., 1988. Disperse Systems. In: Aulton, M.E. (Ed.), Pharmaceutics: The Science of Dosage Form Design. Churchill Livingstone, New York, pp. 81-118. Kobayashi, M., Ishikawa, S., Sameijima, M., 1982. Application of nonlinear viscoelastic analysis by the oscillation method to some pharmaceutical ointments in the Japanese Pharmacopeia. Chem. Pharm. Bull. 30, 4468-4478. Krog, N., 1975. Structures of emulsifier-water mesophases related to emulsion stability. Fette Seifen Anstrichmittel, 77, 267-271. Lashmar, U.T., Beesley, J., 1993. Correlation of rheological properties of an oil in water emulsion with manufacturing procedures and stability. Int. J. Pharm. 91, 59-67. Lee, H.M., Lee, J.W., Park, O.O., 1997. Rheology and dynamics of water-in-oil emulsions under steady and dynamic shear flow. J. Colloid Interface Sci. 185, 297-305. Marriott, C., 1988. Rheology and Flow of Fluids. In: Aulton, M.E. (Ed.), Pharmaceutics: The Science of Dosage Form Design. Churchill Livingstone, New York. pp. 17-37. Martin, A., 1993. Physical Pharmacy. 4th Ed., Lea & Febiger, Philadelphia. Mollet, H., Grubenmann, A., 2001. Microemulsions, Vesicles, and Liposomes. In: Formulation Technology, Emulsions, Suspensions, Solid Forms. Wiley-VCH, Weinheim. pp. 105-122. Morrison, F.A., 2001. Understanding Rheology. Oxford University Press, Inc., New York. Myers, D., 1988. Surfactant Science and Technology. VCH Publishers, Inc., New York. Navarre, M.G., 1941. The Chemistry and Manufacture of Cosmetics. D.Van Nostrand Company, Inc., New York, pp. 187-208. Opawale, F.O., Burgess, D.J., 1998a. Influence of interfacial rheological properties of mixed emulsifier films on the stability of water-in-oil-in-water emulsions. J. Pharm. Pharmacol. 50, 965-973. 56

Opawale, F.O., Burgess, D.J., 1998b. Influence of interfacial properties of lipophilic surfactants on water-in-oil emulsion stability. J. Colloid Interface Sci. 197, 142-150. Pal, R., Bhattacharya, S.N., Rhodes, E., 1986. Flow behaviour of oil-in-water emulsions. Can. J. Chem. Eng. 64, 3-10. Pal, R., 1999a. Yield stress and viscoelastic properties of high internal phase ratio emulsions. Colloid Polym. Sci. 277, 583-588. Pal, R., 1999b. Flow properties of high internal phase ratio oil-in-water emulsions. Appl. Mech. Eng. 4, 363-368. Peltonen, L., Yliruusi, J., 2000. Surface pressure, hysteresis, interfacial tension, and cmc of four sorbitan monoesters at water-air, water-hexane, and hexane-air interfaces. J. Colloid Interface Sci. 277, 1-6. Radebaugh, G.W., Simonelli, A.P., 1983. Phenomenological viscoelasticity of a heterogenous pharmaceutical semisolid. J. Pharm. Sci., 72, 415-422. Rakshit, A.K., Zografi, G., Jalal, I.M., Gunstone, F.D., 1981. Monolayer properties of fatty acids. II. Surface vapor pressure and the free energy of compression. J. Colloid Interface Sci. 80, 466-473. Rieger, M.M., 1986. Emulsions. In: Lachman, L., Lieberman, H.A., Kanig, J.L. (Ed.), The Theory and Practice of Industrial Pharmacy, 3rd Ed., Lea & Febiger, Philadelphia, pp. 502533. Rieger, M.M., 1991. Stability testing of macroemulsions. Cosmet. Toiletries 106, 59-69. Santos Magalhaes, N.S., Cave, G., Seiller, M., Benita, S., 1991. The stability and in vitro release kinetics of a clofibride emulsion. Int. J. Pharm. 76, 225-237. Schott H., 1980. Saturation adsorption at interfaces of surfactant solutions. J. Pharm. Sci. 69, 852-854. Schramm, G., 1994. A Practical Approach to Rheology and Rheometry. Gebrueder HAAKE GmbH, Karlsruhe. Sherman, P., 1964. The flow properties of emulsions. J. Pharm. Pharmacol. 16, 1-25. Sherman, P., 1968. Rheology of emulsions. In: Sherman, P. (Ed.), Emulsion Science. Academic Press, London, pp. 217-347.

57

Shinoda, K., Yoneyama, T., 1980. Evaluation of emulsifier blending. J. Disper. Sci. Tech. 1, 1-12. Sjblom, J. (Ed.), 1996. Emulsions and emulsion stability. Marcel Dekker, Inc., New York. Stokes, R.J., Evans, D.F., 1997. Fundamentals of interfacial engineering. Wiley-VCH, Inc., New York. Tamburic, S., Craig, D.Q.M., Vuleta, G., Milic, J., 1996. An investigation into the use of thermorheology and texture analysis in the evaluation of w/o creams stabilized with a silicone emulsifier. Pharm. Dev. Tech. 1, 299-306. Terrisse, I., Seiller, M., Rabaron, A., Grossiord, J.L., 1993. Rheology: how to characterize and to predict the evolution of w/o/w multiple emulsions. Int. J. Cosmet. Sci. 15, 53-62. Tschoegl, N. W., 1989. The Phenomenological Theory of Linear Viscoelastic Behaviour, An Introduction. Springer-Verlag, Berlin. Vasiljevic, D., Vuleta, G., Dakovic, L.J., Primorac, M., 1994. The influence of emulsifier concentration on the rheological behaviour of w/o/w multiple emulsions. Pharmazie 49, 933934. Virtanen, S., Selkinaho, E., Yliruusi, J. 1993. Effect of the amount of nonionic emulsifier and cooling rate on the rheological properties of some emulsions. Pharm. Tech. Europe, October 1993, 40-50. Weiner, N., 1986. Strategies for formulation and evaluation of emulsions and suspensions: some thermodynamic considerations. Drug Dev. Ind. Pharm. 12, 933-951. Wood, J.H., 1986. Pharmaceutical Rheology. In: Lachman, L., Lieberman, H.A., Kanig, J.L. (Ed.), The Theory and Practice of Industrial Pharmacy. Lea & Febiger, Philadelphia, pp. 123145. Zografi, G., 1982. Physical stability assessment of emulsions and related disperse systems: a critical review. J. Soc. Cosmet. Chem. 33, 345-358.

58