Fat crystallisation at oil–water interfaces

Advances in Colloid and Interface Science 203 (2014) 1–10

Contents lists available at ScienceDirect

Advances in Colloid and Interface Science journal homepage: www.elsevier.com/locate/cis

Fat crystallisation at oil–water interfaces M. Douaire 1, V. di Bari, J.E. Norton, A. Sullo, P. Lillford, I.T. Norton Chemical Engineering, the University of Birmingham, Edgbaston, Birmingham B15 2TT, UK

a r t i c l e

i n f o

Available online 30 October 2013 Keywords: Emulsion Fat crystal Interfacial crystallisation Heterogeneous nucleation Templating

a b s t r a c t This review focuses on recent advances in the understanding of lipid crystallisation at or in the vicinity of an interface in emulsified systems and the consequences regarding stability, structure and thermal behaviour. Amphiphilic molecules such as emulsifiers are preferably adsorbed at the interface. Such molecules are known for their ability to interact with triglycerides under certain conditions. In the same manner that inorganic crystals grown on an organic matrix see their nucleation, morphology and structure controlled by the underlying matrix, recent studies report a templating effect linked to the presence of emulsifiers at the oil/water interface. Emulsifiers affect fat crystallisation and fat crystal behaviour in numerous ways, acting as impurities seeding nucleation and, in some cases, retarding or enhancing polymorphic transitions towards more stable forms. This understanding is of crucial importance for the design of stable structures within emulsions, regardless of whether the system is oil or water continuous. In this paper, crystallisation mechanisms are briefly described, as well as recent technical advances that allow the study of crystallisation and crystal forms. Indeed, the study of the interface and of its effect on lipid crystallisation in emulsions has been limited for a long time by the lack of in-situ investigative techniques. This review also highlights reported interfacial effects in food and pharmaceutical emulsion systems. These effects are strongly linked to the presence of emulsifiers at the interface and their effects on crystallisation kinetics, and crystal morphology and stability. © 2013 Elsevier B.V. All rights reserved.

Contents 1. 2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . Crystallisation, a general phenomenon . . . . . . . . . . . . . . . 2.1. Crystallisation of triacylglycerols (TAGs) . . . . . . . . . . . 2.2. Polymorphism . . . . . . . . . . . . . . . . . . . . . . . 2.3. Nucleation and growth . . . . . . . . . . . . . . . . . . . 3. Recent technical advances allowing the study of interfacial crystallisation 4. Crystallisation at the interface . . . . . . . . . . . . . . . . . . . 4.1. Fat crystals and emulsion stability . . . . . . . . . . . . . . 4.2. Templating nucleation and crystallisation promoters . . . . . . 4.3. Effect on polymorphic form and crystal arrangement . . . . . . 4.3.1. Control of the polymorphic form . . . . . . . . . . 4.3.2. Smaller crystals and mixed crystals . . . . . . . . . 4.3.3. Orientated crystals/crystal arrangement . . . . . . . 5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Thermodynamic aspects of thermal phase transition and the kinetics of crystallisation in lipids are affected by several factors, including their physical state, i.e. the bulk versus emulsified state. Intuitively, some

1

E-mail address: [email protected] (M. Douaire). Tel.: +44 121 414 5284.

0001-8686/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cis.2013.10.022

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

1 2 2 2 2 3 4 4 5 6 6 7 8 8 8

differences in thermal phase transition of lipids are expected to occur in cases where they exist as part of physically heterogeneous systems, such as emulsions. Several parameters can affect interfacial crystallisation and/or crystallisation in emulsions: • Type of emulsion (oil continuous or water continuous); • Type of nucleation process (surface – heterogeneous nucleation, or volume – homogeneous dependent nucleation);

2

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10

• Composition of the lipid phase and molecular interaction with the additive or emulsifier present; • Molecular packing geometry and mobility of surfactant at the interface. It has been long observed that oil contained within a droplet requires more supercooling than the equivalent bulk fat, the most accepted explanation being that these droplets are statistically free of impurities, therefore preventing heterogeneous nucleation. More recently, a difference in crystallisation and/or melting behaviour of oil continuous systems has been highlighted. For example, Alexa and co-workers [1] noticed that the melting temperature of water in oil (W/O) spreads decreases as the κ-carrageenan concentration within the internal aqueous phase increases. It has been suggested that this is due to the destabilisation effect of the κ-carrageenan on the emulsion, which leads to the lower melting point observed. The following paragraphs aim to describe the mechanisms involved in interfacial crystallisation, and the consequences regarding the physical behaviour of emulsion systems. Recent studies have shown that the oil–water interface plays a crucial role in the crystallisation of emulsions, which is of high importance in the design of food emulsion structures, especially as emulsions systems are increasingly designed as tools for controlled delivery [2]. Regardless of the type of emulsion considered (i.e. oil in water (O/W), water in oil (W/O), or double emulsions systems), crystallisation at the interface has a great effect as it will determine the stability of the structure formed. Numerous studies have been carried out in order to explain the mechanisms leading to either improved stability or de-emulsification. These studies aim to either describe the phenomenon or to understand the mechanisms of interfacial nucleation and crystallisation. Interfacial crystallisation of triglycerides (TAG) will, in most cases, happen in the presence of an emulsifier (either protein, monoglyceride, polyglycerol polyricinoleate (PGPR), lecithin, sucrose ester), with each type of emulsifier exhibiting a different effect on the crystallisation behaviour. It is fundamental to understand the mechanisms of heterogeneous nucleation at interfaces, which may be through hydrophobic interactions and templating. In the following sections, we will try to summarise these observed effects in both ‘bulk’ conditions and within an emulsified system, bearing in mind that nucleation (heterogeneous nucleation via either templating effects or prior crystallisation of the emulsifier present at the interface), growth rate (crystallisation kinetics) and crystal morphology are affected by the presence of additives, droplet–droplet interaction (protruding crystals) and interfacial membrane structure. While fat crystallisation in dispersed oil droplets has been extensively reviewed [2–5], here we will highlight recent advances in the understanding of interfacial effects on fat crystallisation in W/O and O/W emulsion systems. The observed effects of the interface will be discussed, leading to concluding remarks on the overall impact of interfacial films on the crystallisation phenomenon in food and pharmaceutical applications.

and R3; for example, oleic, linoleic, lauric, palmitic or stearic) arranged on a glycerol molecule, with TAG species varying in fatty acid chain length (i.e. carbon number, typically between 12 and 24), degree of saturation of the fatty acids (i.e. number and position of double bonds) and arrangement on the glycerol backbone (i.e. positions sn-1, sn-2 and sn-3 [7–10]). Crystallisation is a first-order transition: the process of formation of solid crystals from the liquid state (i.e. melt). When TAGs nucleate the molecules orient in a ‘chair’ (where the fatty acids in sn-1 and sn-2 become the legs, and sn-3 becomes the back of the chair) or an asymmetric ‘tuning fork’ configuration (where the fatty acid chains in positions sn-1 and sn-3 point in one direction, and the fatty acid in sn-2 points in the opposite direction) [11]. During TAG crystallisation, TAG molecules stack in pairs that self-assemble into lamellae, which in turn stack into crystalline nanoplatelets, that aggregate to form clusters (primary crystal particles), that then pack in an arbitrary manner to form flocs, until a three-dimensional network is created [8,11–13]. 2.2. Polymorphism The TAG molecule can crystallise into different crystalline forms in the crystal lattice (i.e. conformation and arrangement), depending on processing conditions (e.g. cooling rate and shear), a phenomenon termed polymorphism. As such, rapid cooling of liquid fat results in the formation of a diffuse crystalline phase (low-energy polymorph), whereas slower cooling means that the molecules have time to organise into lamellae to form consistent, three-dimensional crystals. Polymorphic behaviour, such as melting point, is affected by the chain length of the fatty acids within the TAG [7,10]. Often the polymorphic forms of fats are classified into three categories, α, β′, and β, in increasing order of stability (due to the density of hydrocarbon chain packing), melting point and latent heat of fusion. The packing of the hydrocarbon chains is different for the different polymorphs. This can be explained in a simplistic way, whereby α has a disorganised, least dense hexagonal (H) subcell structure, β′ an intermediate packed orthorhombic perpendicular (O⊥) configuration, and β a tightly packed triclinic parallel (T//) subcell structure [14]. The main polymorphic forms are illustrated in Fig. 1 [15]. The polymorphic form of fat can be measured using X-ray diffraction patterns (short spacings), and inferred using differential scanning calorimetry. Lipids exhibit monotropic polymorphism (i.e. they can exist in multiple forms, but only one form is stable at all temperatures and pressures, though different polymorphs can coexist). Unstable forms are the first to crystallise in a subcooled fat. Subsequent transformation of unstable polymorphs into more stable forms occurs over time until, eventually, the most stable polymorph for a given lipid is reached. The difference in Gibb's free energy (G) between polymorphs is the driving force for this transformation, which is defined as:

2. Crystallisation, a general phenomenon G ¼ H−TS In order to understand crystallisation within emulsions (both in the bulk and at the interface), it is important to briefly cover the key features of crystallisation in bulk fats, including polymorphism, nucleation and growth, crystallisation in mixed systems, and the effect of temperature and shear. In food systems, fat crystal number, size (both the mean size and the distribution) and polymorph all affect the physical and textural properties, and subsequent sensory properties (e.g. appearance, mouthfeel or flavour release), of the final product. 2.1. Crystallisation of triacylglycerols (TAGs) Fats and oils are complex mixtures of triacylglycerols (TAGs) (which typically make up approximately 98% of the mixture), and more polar lipids like diacylglycerols (DAGs), monoacylglycerols (MAGs), free fatty acids (FFAs), phospholipids, glycolipids, sterols, and other minor components [6] TAGs are composed of three fatty acyls/acids (R1, R2

where H is enthalpy, S is entropy, and T is temperature [10,14]. Values of Gibbs free energy are highest in α, intermediate in β′ and lowest in β. Fatty acid chain length affects the rate of transformation, with faster transformation for TAGs with short-chain fatty acids. This transformation is also shear and temperature dependent, which can be utilised to produce particular, desirable, polymorphs (for example, tempering of cocoa butter/chocolate). Interpolymorphic transitions are unidirectional (i.e. β cannot transform to β′ and β cannot transform to α), without returning to system to the melt [10]. 2.3. Nucleation and growth Crystallisation is a two stage process, involving the formation of nuclei followed by crystal growth. Crystallisation occurs providing that the phase is supersaturated (i.e. the concentration of dissolved species is

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10

3

similar melting points; the melting point of the blend is lower than the melting point of the individual components) or 3) compound crystals (synergistic compatibility between two TAGs, improving packing ability) [14]. In TAG mixtures phase behaviour is affected by chain length, degree of saturation, nature of any double bonds, and arrangement of the fatty acids on the glycerol backbone. Cooling rate has a dramatic effect on crystallisation, affecting both crystallisation rate (as a result of time to rearrange into a crystal lattice), and nucleation rate, which governs crystal size; rapid cooling to a low temperature promotes a higher nucleation rate, resulting in many small crystals. Shear also affects nucleation and crystal growth, due to mechanical disturbance that supplies energy to overcome the energy barrier for nucleation, and promotes secondary nucleation. Post-crystallisation processes include the bridging between crystals to form a network. 3. Recent technical advances allowing the study of interfacial crystallisation

Fig. 1. Structure models and Gibbs energy (G)–temperature relationship of three polymorphs of PPP, reproduced from Sato [15].

greater than the equilibrium concentration, i.e. the concentration where molecules are in the equilibrium state). This can be achieved by cooling (i.e. cooling below the equilibrium point) until a free-energy barrier is overcome and a stable nuclei is formed. When nucleation does occur, there is a release of energy (latent heat of fusion) as the molecules assume the lower energy state in the crystal lattice. Nucleation can either be as primary (homogeneous or heterogeneous) or secondary nucleation. Primary nucleation occurs in the absence of crystals (i.e. straight from liquid state to crystal formation), whilst secondary nucleation occurs if crystals of the same species are already present. Homogeneous nucleation occurs when there is an accumulation of molecules in the liquid state until a nucleus forms, whereas in heterogeneous nucleation foreign nucleating sites (i.e. impurities) catalyse nucleation (reducing the free-energy required for nucleation). Secondary nucleation occurs when new nuclei are formed in the presence of existing crystals, typically due to fragments of growing crystals that are chipped off and act as nuclei, collisions between two crystals, or between a crystal and a surface, such as a stirrer or vessel walls. In industrial settings nucleation can be promoted with the addition of crystal ‘seeds’. Once nuclei have formed they grow by the incorporation of other TAG molecules from the liquid phase into the surface of the crystal lattice. In order for growth to occur, molecules must migrate to the surface of the crystal (diffusion), and orient to an appropriate site for incorporation into the lattice (deposition). This leads to a release of latent heat; this energy must be transported away from the crystal surface to prevent temperature increase, or no further growth can occur. Crystal growth continues, typically in an exponential manner, until there is a phase equilibrium or the entire system is crystallised. The morphology of the crystal is determined by the composition of the mixture, and by growth rate. The structure is modified as the size of the crystals increases overtime due to Ostwald ripening: small crystals dissolve in solution, while larger crystals grow as they are energetically favoured. Whilst in the liquid state TAGs are miscible, on crystallisation the different TAG species can come together to form 1) solid-solution mixtures (when the TAGs are very similar in melting temperature, molecular volume and polymorphism; the melting point of the mixture is between that of the pure components), 2) eutectic mixtures (when the TAGs differ in chain-length, molecular volume, shape and polymorph, but have

The role of the interface and the physical confinement on the nucleation and growth of crystals becomes particularly crucial in micro-scale systems, such as emulsions, where the ratio of interface to bulk is extremely high [2,5]. The high scientific interest in this area of research has drawn the attention to the limitation in experimental methods for the direct probing of local structure and interface dynamics. The majority of the published studies are conducted using “bulk techniques”, where all emulsion droplets are subjected to the measurement simultaneously. The range of techniques include: nuclear magnetic resonance [16–18], differential scanning calorimetry [19,20], X-ray diffraction [21–24] and ultrasonic velocity [25–27]. Amongst those techniques, differential scanning calorimetry (DSC) has been demonstrated over the years to be particularly suitable to provide an insight into the thermodynamics of fat phase transitions, either in bulk [28–30] or in emulsion systems [19,20]. In general, a change in physical state of a material (e.g. melting, crystallisation), or a transition from one crystalline form to another (e.g. polymorphic transition) is accompanied by the absorption (endothermic) or release (exothermic) of heat which is measured by the DSC (heat flow) as a function of time and temperature. A detailed description of the theory can be found in several published texts [31]. The enthalpic change can be measured as the area under the peak whilst the direction of the peak indicates whether the thermal event is endothermic or exothermic. The peak shape and temperature give indication about the polymorphic form of a crystal. According to classical crystallisation theories [32], formation of less perfect crystals occurs as the degree of supercooling increases, which results in higher melting temperatures peak. On the other hand, measurements under isothermal conditions can be indicative of the nature of the crystallisation process. For instance, the crystal growth mechanism can be described using the well known Avrami equation [33], while the induction time (inversely proportional to the theoretical nucleation rate) might be used to evaluate the activation free energy for a stable nucleus using the Fisher–Turnbull equation [34]. Recently, Foubert and co-workers [35] proposed a valuable alternative method (stop-and-return) based on the determination of the melting profiles at different moments in time during the experiment. It provides more information about the crystallisation mechanism when real-time X-ray diffraction is not readily available. Using this methodology Frederick and co-workers [19] reported on the two-step crystallisation mechanism for either bulk or the emulsified milk fat. However, great care should be taken when comparing crystallisation occurring in the bulk and in emulsified system. For instance, in O/W emulsions the heat transfer to or from the oil droplets occurs via the aqueous phase almost instantaneously. Conversely, aggregation and sintering of bulk crystals prevent any rapid mixing of the highly viscous mass, favouring the development of a temperature gradient within the bulk. A disadvantage of this technique is that it does not measure the presence of a crystal form directly, which renders the study of polymorphism rather challenging. Moreover, if

4

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10

several processes occur simultaneously, a superposition of heat effects is measured which cannot be readily de-convoluted [22]. The use of X-ray diffraction in combination with DSC makes it possible to disentangle the sequential thermal changes and to overcome the problem mentioned above [22,36]. X-ray diffraction (XRD) is generally used to provide information on the crystal lattice geometry and to distinguish between polymorphic forms. In general, X-rays are diffracted by the crystalline structure at a determined angle by the planes of the atoms in the structure; the set of peaks present in an X-ray diffractogram corresponds to the distance between the crystallographic planes. The main features of the molecular arrangement in layers can be obtained by the diffraction at small angles, called long spacing. Instead short spacing, diffraction at high angles, allows for determination of the polymorphs α, β′, and β [37]. New opportunities are coming from scientific fields where morphological characterisation tools are routinely used for the analysis of nanostructure materials in the form of thin films. To this end Grazing Incidence Small-Angle X-ray scattering (GISAXS) has emerged as a powerful non destructive technique to measure the molecular ordering with subnanometer resolution of absorbed monolayers on solid surface as well as at the liquid/liquid interfaces [38–40] The problems of small diffracting volume or low diffracted intensities compared to the substrate and background usually encountered during the analysis of thin film structures are overcome by increasing the path length of the incident X-ray beam through the film [41]. The “surface sensitivity” necessary for the investigation of thin films with thicknesses down to the sub-monomolecular range, is provided by the use of synchrotron radiation at very low incident angles of the beam (grazing incidence geometry) to allow small penetration depths (only the top or less into the surface) and to maximize the signal intensity from the thin layer at the surface of the material. [42]. A detailed review of the theoretical basis of GISAXS and the X-ray refraction reflection effects at low incidence angle of the beam with respect to the surface has been recently published by Renaud et al. [42]. Using this technique Tikhonhov et al. [40] have studied the ordering presence of solid close-packed surfactant assemblies at the water–hexane liquid–liquid interface where conformational changes of the tail group from disordered to ordered monolayers are strongly dependent on the fine chemistry of head group substitution. It is expected that further development of such techniques and the adequate sample preparation will provide such levels of structural information of fat crystals at the oil–water interface. Recent studies on the polymorphism and crystallisation in emulsion systems make use of very high energy beam (Synchrotron), which enables one to take a complete and detailed measurement of a sample considerably quickly, allowing a time resolved type of experiment [23,43]. Ollivon and co-workers [44] coupled a DSC and X-ray diffraction in a single apparatus, which allows simultaneous measurement on the same sample. This technique has been successfully employed to investigate the effect of droplet size [21] presence of additives [36] and surface composition [22] on crystallisation phenomenon in emulsified systems. It is worth mentioning that the collection of clear data with good signal to noise ratio is still difficult since it is strongly dependent on the concentration of the sample (dilution effect) and the scanning rates [22]. Moreover, monitoring the process of crystallisation in a single droplet or at the o/w interface is not possible. To overcome this problem Sato's group used a scanning X-ray microbeam technique composed of an ultra intense X-ray source combined with a two-dimensional detector (X-ray CCD detector) [45]. “(2D) X-ray diffraction maps” are constructed collecting two-dimensional X-ray data over a specific area of the sample. The use of synchrotron allows finer resolution, in this case 5 × 5 μm2. Quantitative microscopic information about the local orientation of crystal at different points inside a droplet, of about 30 μm, is now possible. Shinohara and co-workers [45] used this technique to study interfacial heterogeneous crystallisation of n-hexadecane at the O/W interface, where in the presence of hydrophobic emulsifier the chain axis of the crystallised n-hexadecane orientates perpendicular to the oil–water interface. Similarly, [46] reported on the orientation of the crystal lamellar

plane, parallel to the direction of the W/O interface and the promotion of fat crystallisation by the presence of monoglycerides. Although the precise details of the polymorphism at the interface await more advances in the technique, the state-of-the-art described above clearly opens exciting new vistas in the quantitative and qualitative description of crystallisation at the interface. 4. Crystallisation at the interface 4.1. Fat crystals and emulsion stability Fat crystals at an oil–water interface can have two opposite effects on emulsion stability. In water continuous systems, protruding crystals at the interface usually lead to emulsion destabilisation and partial coalescence [47]. Stiffer crystals are more likely to be involved in destabilisation of water continuous emulsions, as they are less adaptable to curvature angle of the interface [48]. This phenomenon has been extensively reviewed (see for example [3]) due to its crucial importance in certain food process, e.g. ice cream or butter. In oil continuous emulsions, fat crystals usually enhance stability by increasing interfacial viscosity [49,50], acting as Pickering particles [3,4,51–53] or even forming a crystalline shell of sintered crystals around water droplet [52,54]. The mechanism of formation of these shells is currently not well understood. Fat crystals also facilitate the emulsification process by lowering the interfacial tension [55]. The formation of a crystalline interfacial film occurs at the oil–water interface when the limit of association of the emulsifier in the bulk oil phase has been reached [55]. Along with formation of the crystalline film (which also depends on emulsifier concentration and chain length), a decrease in the interfacial tension occurs upon cooling. Although the system studied was not an emulsion, Krog and Larsson [55] introduced the concept of interfacial crystallisation upon emulsification, which implies the formation of “surface-active crystals” at the interface. The tendency of a fat crystal (or any Pickering particle) to sit at an interface is called wettability, measured in terms of contact angles between the crystal and the hydrophobic and hydrophilic liquid surfaces (see Fig. 2). The curvature of the interface is controlled by the contact angle (θ), thus, for oil and water systems, the type of emulsion stabilised. Fat crystals tend to favour the stabilisation of oil continuous emulsions. Their size and shape determine their effectiveness as stabilisers. Needle like crystals lead to rapid destabilisation [56] whereas small platelets improve the stability of W/O emulsions [57,58] and smaller crystals lead to improved stability due to better surface coverage [4,53,59]. Furthermore, several authors have shown that, with regard to Pickering stabilisation of emulsions, in-situ crystallisation is more effective than pre-crystallisation of fat particles [3,60,61].

Fig. 2. Contact angle of a particle at an interface.

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10

4.2. Templating nucleation and crystallisation promoters Nucleation can be controlled by surface or bulk conditions [62]. In the case of homogeneous nucleation, a high degree of supercooling, long residence time or supersaturation is required to form crystal nuclei. Hence, there is a difference in crystallisation/melting behaviour and solid fat content between bulk and emulsified fats [63]; within oil droplets, impurities that could induce heterogeneous nucleation are virtually dissolved, being less than one nucleus per droplet. In the case of heterogeneous nucleation, however, impurities from the melt will act as a starting point. Additives are sometimes included in the melt as crystallisation promoters or inhibiters. In emulsions, the arrangement of emulsifier molecules adsorbed at the oil/water interface can sometimes create a matrix which templates the interfacial heterogeneous nucleation of the TAG present in the oil phase. These organised structures lower the activation energy of nucleation (nuclei stabilised by specific interactions with the matrix, due to compatibility between surfaces and molecular ordering between matching TAG and emulsifier tails). For such a templating effect to occur, the hydrophobic tail of the emulsifier and the TAG should exhibit some level of structural similarity. In this case, the created interfacial membrane modifies the heterogeneous nucleation process [64], crystallisation kinetics [36,65–67] and secondary nucleation process (which has been shown to be mediated by droplet collision in Tween20 stabilised O/W emulsions, for example) [68]. Molecular interaction can occur when chains match in length [45] and structure [69], thus exhibiting enough structural similarities with the oil to induce nucleation [5,69]. This compatibility between emulsifier and TAG has also been highlighted in dry oil phase systems [70–72] for example). However, Gülseren and Coupland [73] noticed a similar effect for a C16 emulsifier tail combined with a C18 alkane, suggesting that the match does not have to be perfect. The specific effects of the type and molecular structure of the emulsifier on the crystallisation behaviour in emulsions have previously been highlighted for some molecules [69,74,75]. Frederick and co-workers [18] studied the effect of monoglyceride chain length and structure on milk fat crystallisation in recombined cream. They showed that stearic acid acts as a crystallisation promoter that also enhances polymorphic

5

transition whereas unsaturated monoglyceride (oleic acid) does not affect crystallisation kinetic not polymorphic transition. It is also worth noting that Relkin and co-workers [76] showed that caseins lower crystallisation temperature and affect the crystallisation profile of fat droplets compared to whey protein isolate. A templating effect has been reported in both oil and water continuous systems. Kaneko and co-workers [66] reported an increase in the nucleation rate of n-hexadecane with sucrose ester (SE) in water continuous emulsions compared to the equivalent oil phase, suggesting a templating effect of the SE. As an example, Arima and co-workers [77] showed that the adsorption of hydrophobic additive (in this case sucrose oligoesters containing palmitic acid) at the palm mid fraction (PMF, a fraction of palm oil rich in Palmitic–Oleic–Palmitic triglyceride)–water interface increases the crystallisation temperature of the droplets (see Fig. 3). Sakamoto and co-workers [78] studied the crystallisation of emulsified palm mid fraction in the presence of polyglycerol fatty acid esters. Polyglycerol was esterified with saturated fatty acid (palmitic, stearic or behenic acid). They noticed an increase in crystallisation temperature (Tc) with the addition of as little as 0.1% of emulsifier. The increase in Tc was enhanced for high degrees of esterification and increased fatty acid chain length. They concluded that the emulsifier enhanced interfacial heterogeneous nucleation when present, explaining the stronger effect of long chain fatty acid. However, further addition of polyglycerol 10 dodecabehenate had little effect. As one can assume that addition of emulsifier would only increase its concentration in the bulk (0.1% is usually enough for complete surface coverage of microemulsion), this would suggest the main effect in this case was heterogeneous interfacial nucleation. DAGs have also been shown to template for nucleation of n-hexadecane droplets, with the templating effect increasing with the length of the C chain of the fatty acid moieties [79]. Kalnin and co-workers [22] proposed a slightly different explanation to justify the effect of emulsifier on fat crystallisation. Using time-resolved synchrotron X-ray diffraction (XRDT) coupled with DSC, they studied the effect of propylene glycol monostearate (PMGS) on crystallisation and polymorphism in palm oil droplets dispersed in an aqueous medium. The results showed that PGMS, even at a low concentration (range from 0 to 0.2%), was able to increase the temperatures of

Fig. 3. Interfacial heterogeneous nucleation mechanism (adapted from Wassel et al., Ueno et al. and Arima et al. [46,64,76]) and melting point depression (adapted from Povey et al. [93]).

6

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10

the onset, maximum and offset of crystallisation, and decrease the enthalpy of crystallisation by 20%. It is suggested that PGMS interacts with shorter chain TAG and, in the presence of water, forms a lamellar structure at the oil–water interface, forming liquid crystals. The solubilisation of PGMS by liquid oil at the interface results in a locally increased concentration of long chain, high melting TAG: the presence of an interfacial template helps the enrichment of the oil phase on high melting TAG by ‘entrapping’ low melting TAG. The presence of liquid crystals has also been invoked [80] to explain the slow release of limonene from O/W nano-emulsion stabilised by Tween 20 and whey protein isolate in the presence of glycerol mono stearate. Fewer studies have been carried out on W/O systems. However, there is evidence of templating effects described in the studies of Ghosh and Rousseau [81] and Ueno's group [46,82]. The latter showed that the presence of monobehenodylglycerol (MB), a high melting monoglyceride, at the interface provides a template for the bulk TAGs, promoting fat crystallisation at the interface (see Fig. 3) as well as in the continuous phase, and the formation of small crystals. Their findings were supported by DSC analysis on cooling, which showed a split (i.e. double) exothermic peak for PGPR stabilised emulsions with added MB as opposed to a single peak for emulsions containing PGPR only. However, the somehow limited adsorption of MB at the interface compared to PGPR lessens the effect of MB on interfacial crystallisation. It has been noticed that crystallisation of hydrogenated canola oil (HCO) in emulsions stabilised by PGPR (0.25 wt.%) and glycerol monooleate (GMO) (2 wt.%) started at the interface, with continuous phase crystallisation being delayed [81]. They attributed this to facilitated HCO nucleation at the interface in the presence of GMO, a feature that does not appear in PGPR stabilised emulsions. Similar templating effects of GMO in stearine containing oil droplets have also been suggested [83]. It is interesting to note that there is no evidence of effect of GMO on crystallisation behaviour of palm mid fraction [84] in the bulk state, the templating properties of GMO only being given by their position at an oil/water interface and the molecular compatibility with the TAG present. These concepts are illustrated in Fig. 2. 4.3. Effect on polymorphic form and crystal arrangement 4.3.1. Control of the polymorphic form The control of the polymorphic form can become important as the transition to a more stable form can lead to needle-shaped crystals that protrude through the interface, with their eventual bridging leading to emulsion destabilisation [3]. Furthermore, polymorphs do not all have the same affinity for the interface, as α form crystals are slightly more hydrophilic than β and β′ [57,85]. Finally, the polymorphic form of fat crystals has been shown to have an effect on interfacial viscosity, which in turn affects the stability of protein stabilised emulsions. The presence of fat crystals at the interface of oil and an aqueous solution of proteins increases interfacial shear viscosity, with the effect being more pronounced for β′ crystals [49]. The authors suggested that the protein layer binds with the fat crystals, leading to more crystal–crystal interaction when sheared, as this behaviour was not exhibited in similar experiments performed with small molecular surfactant instead of proteins. The organised structure provided by the arrangement of amphiphilic molecules at the interface has also been shown to induce a different crystal structure compared to bulk crystals [64]. Using small and wide angle X-ray diffraction coupled with DSC, Ueno and co-workers observed the formation of triclinic crystals in oil droplets without high melting additives, as well as in bulk oil phase, regardless of the presence of these additives. However, when crystallised in oil droplets in the presence of additives, pseudohexagonal and orthorhombic crystals were formed instead, indicating that particular interfacial organisation can lead to specific crystal structures. As an example, sucrose oligoesters with stearic moieties have been shown to induce crystallisation in the β′ form of PKO [67] and PMF [36] for O/W emulsions in the presence

of Tween 20 (see Fig. 3). Frederick and co-workers [18] showed that polymorphic transition of milk fat in recombined cream was enhanced when long chain saturated monoglycerides were added but no effect with saturated ones, due to constraints in the microstructural arrangement of the fat within droplets in the presence of certain category of monoglycerides. Arima and co-workers [86] observed that the addition of sucrose esters (SE) retards crystallisation induced destabilisation of O/W emulsions stabilised by Tween 20. They studied the effect of addition of hydrophobic SE and highly hydrophilic SE containing palmitic acid moieties on PMF droplet destabilisation. Their results showed that the addition of hydrophobic SE, despite increasing the crystallisation temperature of the droplets, retards crystallisation induced destabilisation. This is due to the enhancement of interfacial heterogeneous crystallisation, which in turn induces the formation of smaller crystals, as mentioned previously. Furthermore, through optical observation, DSC and X-ray studies, they noticed a synergistic effect between the two additives as the hydrophilic SE retards polymorphic transition, thus preventing the formation of detrimental needle shaped crystals. Further studies by the same group [77] compared polymorphic transitions in O/W emulsions stabilised by Tween 80. They showed that the addition of hydrophobic SE also retards polymorphic transition compared to emulsions made in its absence. It is interesting to note here that these additives have also been reported to slow down polymorphic transitions in dry systems (Smith et al. 2012). Kalnin and co-workers [22] showed that the addition of PGMS reduces the rate of polymorphic transition from α to β′ form. The influence of emulsifier type on polymorphic form of dispersed fat has also been observed for nano emulsions, where the droplet size and the consequently high curvature angle further restrict molecular movement. Nik and co-workers [87] showed that polymorphism in nano droplets of canola stearin was affected by the type of emulsifier used, i.e. β′ coexists with β in the presence of Tween 20 whereas only β occurs in Poloxamer stabilised emulsions. Similarly, Bunje [88] and co-workers showed that crystal structure of triglycerides in solid lipid nanoparticles obtained by melt-homogenisation was affected by the type of emulsifier used and the cooling rate. In glycoholate-stabilised particles, the metastable α form is obtained regardless of the cooling rate applied, but the layers of triglycerides remain less ordered for high cooling rates, indicating an interaction between the emulsifier used and the ability of triglycerides to form an ordered structure. These studies not only show that physical constraints restrict movement, but also that fat crystals and the emulsifier-crystal arrangement within a droplet can slow the molecular rearrangement during polymorphic transitions. These interactions are expected to exert a strong influence on physical properties of solid lipid nanoparticles obtained by a melt-homogenisation process, i.e. via the formation of a nano-emulsion. A less ordered crystal arrangement in solid lipid nanoparticles, despite leading to somewhat lower thermal stability, allows higher drug loading capacity and triglyceride crystal destructuration by emulsifiers and additives in solid lipid nanoparticles, which is of crucial interest in the field of drug delivery. Helgason and co-workers [89] showed that surfactant coverage influences the crystal structure of solid lipid nanoparticles, evidenced by the melting profile of tripalmitin nanoparticles. Additional melting peaks appear as Tween 20 concentration increases, indicating that the emulsifier interferes with the crystal structure. Another example of interfacial interaction between triglyceride and emulsifier in nano emulsions and their resulting solid nanoparticle is provided by the work of Jenning and co-workers [90]. They reported the creation of solid lipid nanoparticles that consist of a “shell” of crystalline long chain saturated monoglyceride (monoglyceryl behenate) around liquid oil, which has considerably reduced movements due to its interactions with the immobilised interfacial crystalline monoglyceride. For oil-continuous emulsions, no clear trend between additive and their effect on polymorphic form can be deducted (see table 1). Wassel and co-workers [46] reported a lack of any noticeable effect of interfacial composition on polymorphic form. SAXD and WAXD patterns of W/O

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10 Table 1 Effect of selected emulsifiers on polymorphism of the continuous fat phase. Authors

Continuous fat phase

Emulsifiers

Wassel et al. [46]

interesterified: palm stearin/ lauric kernel/liquid rapeseed oil hydrogenated canola oil

PGPR and No effect PGPR + monobehenodylglycerol

PGPR and PGPR + glycerol monooleate

No effect

β-tending palm oil based fat blend

Glycerol monooleate, Glycerol monopalmitate

Monoglycerides induce polymorphic transitions in water in oil system compared to dry fat system

Gosh and Rousseau [81] Shiota et al. [91]

Effect on polymorphic transition

emulsions with and without monobehenodylglycerol (MB) were compared, with no difference being observed. They, however, explained the presence of β form by the long storage period, thus not excluding the possibility that the polymorphic form had been different at an earlier stage. Upon second cooling of the sample after melting, SAXD patterns for the emulsion without PGPR showed the presence of α crystals, whereas it was assumed that a low temperature sub α form was present in systems containing MB. In the same manner, Ghosh and Rousseau [81] observed the same polymorph of hydrogenated canola oil (HCO) in W/O emulsions containing PGPR with and without GMO at the interface. On the contrary, Shiota and co-workers [91] noticed faster polymorphic transition in W/O emulsion than in the bulk fat phase. A β-tending palm oil based fat blend was emulsified in the presence of monoglycerides of various chain lengths and degree of saturation. Transition to the stable β form was faster for emulsions containing GMO or glycerol monopalmitate than for their corresponding non emulsified fat phases. Furthermore, the transition was faster for emulsions containing saturated monoglyceride. The authors argued that acyl–acyl interactions between monoglycerides present at the interface and TAGs from the bulk phase induce these polymorphic transitions as the dry preparation does not exhibit the most stable β form. In an earlier study, Aronhime and co-workers [92] showed that in dry systems, liquid emulsifiers tend to favour polymorphic transformations, most likely due to their lack of compatibility with high melting TAG, which enhances mobility of the latter, and therefore, the freedom to reorganise. The few studies reviewed here do not show the same trend (i.e. emulsifiers that are liquid at room temperature, such as GMO or Tweens do not favour polymorphic transitions), highlighting the difficulty in predicting TAG – surfactant compatibility in emulsion systems from the properties of the mixtures obtained in bulk. Furthermore, in most cases a multicomponent system has to be considered and the emulsifier present should not be neglected. For example, the addition of sucrose oligoesters containing palmitic acid moieties in PMF emulsions stabilised by Tween 20 retards the formation of β′ crystals [36], whereas this transition is favoured when added to emulsions stabilised by Tween 80 [77]. This suggests possible synergetic/antagonistic effect between emulsifier and additive. 4.3.2. Smaller crystals and mixed crystals Increased nucleation (i.e. where there is the same amount of crystallisation material but more starting points) usually leads to smaller crystals. It is, therefore, not surprising that studies report the appearance of smaller crystals when interfacial crystallisation is promoted by additives or high melting emulsifiers. This consequence of interfacial facilitated nucleation has been observed in both O/W [86] and W/O systems. Wassel and co-workers [46] observed a reduction in crystal

7

size for W/O emulsions stabilised by PGPR and MB compared with those stabilised by PGPR only. It has been shown, for water in mineral oil emulsions containing glycerol monooleate as an emulsifier and wax crystals, that stability is enhanced when the whole system is crystallised post-processing (i.e. post-emulsification) [61]. For this system, crystals show affinity for the interface in both pre- and post-crystallised samples. The somewhat increased stability obtained with the postcrystallisation method might be explained by the mechanism of interfacial crystallisation, which produced more nuclei, thus reducing crystal size, and crystal surface modification due to the presence of emulsifier. Gülseren and Coupland [73] studied the effect of emulsifier type on surface melting of emulsifier alkanes. Whilst emulsifier type had no effect on C20 alkanes, Tween 40 and Brij 58 enhanced nucleation of emulsified C18 alkanes; resulting in two melting peaks, i.e. a portion affected by surfactant and a ‘pure’ portion. They also studied the effect of droplet size on crystallisation and melting and reported an increase in melting peak and onset with increasing droplet size. The authors explained this as being a result of the decrease in surface area. The same group also studied the effect of surfactant type (Tween 20 vs sodium caseinate), oil droplet size and chain length on the phase transition of even-numbered emulsified n-alkanes (C16, C18, C20) using microcalorimetry and ultrasonic attenuation measurements [74]. Upon melting, the peak temperature was affected mainly by the droplets size and by the alkane chain length; an increase in the droplet size produced a higher peak occurring over a narrower temperature range while the use of longer chain length alkanes produced a shift to higher temperatures. The emulsifiers significantly affected the shape of the melting peak, as Tween 20 stabilised emulsions showed a bimodal peak while a unimodal peak was found for the sodium caseinate emulsions. The authors stated that the difference in the peak shape reflected difference in the droplets structure. The alkyl chain of a Tween 20 molecule, whilst stabilising the emulsion interface, is able to interact with the alkane chains, and therefore, form mixed crystals at the vicinity of the interface. Surface mixed-crystals will melt first, followed by the melting of the pure crystals located in the core, thus producing a bimodal peak. The melting of the protein stabilised emulsions produced a sharp single peak because the lack of an alkyl chain in the protein structure resulted in the absence of mixed crystals. A similar behaviour was observed using ultrasonic attenuation measurements. On melting, Tween 20 stabilised droplets had a higher excess attenuation compared to the protein stabilised emulsions, with smaller droplets increasing the attenuation. On re-crystallisation, the effect of the emulsifier type was less pronounced, as shown by both the microcalorimetry and ultrasonic attenuation measurements. In fact, although the protein stabilised emulsions crystallised at a temperature about 1 °C higher than the Tween 20 based emulsions, the peaks were mostly unimodal. Furthermore, the authors reported the existence of solid–solid phase transition involving the formation of a “rotator crystal phase”, during both melting and crystallisation. This phase is an intermediate between a liquid and a solid crystalline phase. The analysis of calorimetric data revealed that the peak associated with this transition was higher for emulsions containing Tween 20, the authors suggested a correlation between the phase transition and an interaction of the alkanes with the hydrophobic portion of the emulsifier. Although it could be assumed that the location of this phase was in the vicinity of the interface, no further information was provided on this. This phenomenon had previously been shown by Povey and co-workers [93], explaining phase changes at the surface of the droplets by the formation of a solid monolayer formed of surfactant and dispersed alkane phase. Thus, they explain melting point depression at the surface of nano oil droplets by dissolution of about 10% of the oil present in the lauric acid chain of the Tween 20 present at the interface, therefore forming mixed crystals (see Fig. 4). It is also thought that curvature angle may affect crystal growth [5]. For miniemulsions (100 nm b droplet size b 1 μm), it has been demonstrated that the crystallisation rate and melting temperature increase with the droplet size [94], with these findings being attributed to a change in the curvature angle.

8

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10

crystal, acting as an impurity that prevents the formation of a fat crystal network, i.e. it limits further crystal growth once incorporated into a growing crystal. Finally, phospholipids have been shown to interact with crystallising TAG [100] and lecithins exhibit some structural similarities with DAGs, being built on a glycerol backbone whose third glycerol hydroxyl group is esterified with a phosphate group. Furthermore, and saturated lecithins have been reported to increase stability of metastable forms [101] and to slow the sintering process between crystals by adsorbing to the crystal surface [58,97]. It is important to note here that these effects are probably incurred by the incorporation of lecithins into the growing crystal, as [102] observed gel formation of solid lipid nanoparticles when platelets are not fully covered by phospholipid. 5. Conclusion

Fig. 4. Melting point depression zone, adapted from Povey et al. [93].

4.3.3. Orientated crystals/crystal arrangement Crystal arrangement also has a dramatic effect on emulsion stability, leading to the formation of shells [54], cocoons [81] or liquid crystal template [80]. Long chain saturated esterified fatty acids (glycerol and propylene glycol monoesters) are known to create strong crystalline films at an oil/water interface [95,96]. Crystal morphology (needle, platelet, or spherulite) affects their interfacial properties and can either enhance or limit such arrangements, as do the presence of additives and emulsifiers. The templating effect of hydrophobic additives described above also leads to specific orientation of the crystals and influences the arrangement of fat crystals at the interface. The carbon tail of the emulsifier is orientated normal to the interface [45]. As a consequence, crystal growth results in the formation of crystal platelets parallel to this interface, as observed through Synchrotron radiation microbeam X-ray diffraction (SR-μ-XRD) for O/W emulsions [77] and for W/O systems [46,82]. The presence of a fat crystalline interface (called “shell”) in waterin-oil emulsions containing mono- and diglycerides used as emulsifiers has been visualised using cryo-scanning electron microscopy [54]. It was suggested that sub-micron sized crystals consisting of emulsifier molecules were formed under fast cooling and, if absorbed at the w/o interface, were acting as Pickering particles. These crystals were then driving the sintering of a rigid crystalline shell. Furthermore, by measuring the diffusion of NaCl at different temperatures, the authors showed that the rigid interface acts as an efficient mechanical barrier between two aqueous environments. Optimal Pickering stabilisation was achieved by using 0.5% of monoglyceride and 2.0% of tripalmitin, which also gave the smoothest sintered shell, probably as a result of a higher degree of interfacial and bulk crystallisation. The choice of emulsifier can also affect the propensity of crystals to associate. For example, GMO enhances bridging between fat crystals [97] and crystal growth [98]. Conversely, PGPR, is known to adsorb to fat crystal nuclei, favours β′ form instead of α, and prevents crystal growth and aggregation [56]. [81] argue that PGPR inhibits interfacial adhesion of TAGs due to lack of compatibility between the molecules. For water continuous emulsions, structure formation between fat crystals is prevented by the addition of polyglycerol fatty acid full ester [99] due to its crystal modifying properties. The authors suggest that this additive acts as an analogue of TAG and adsorbs to the surface of the fat

TAG crystallisation is affected by the presence of an oil–water interface. The presence of additives generally acts as impurities inducing and/or templating nucleation. Interactions between growing fat crystals and emulsifier templates have to be expected due the presence of an interfacial layer of emulsifier. These interactions are stronger if some degree of compatibility exists between chains. However, for a high degree of supercooling, these effects are usually lessened. Unfortunately, very few studies have compared the effects of additives in wet and dry systems. Emulsifier dissolved in the fat phase will form micelles, whereas in emulsified systems, at least part of them will be organised/structured at the interface. For most dry mixtures of TAG/additives, these interactions have a significant effect on nucleation, but very little effect on growth. Therefore, if there is a constant growth rate but more nuclei, smaller crystals will result, a phenomenon which is of interest for a wide range of Pickering stabilised emulsions. Other additives (phospholipids) have crucial effects on crystal growth rate and shape, and this ought to be studied and used for the formation of fat crystal stabilised interfaces. These advances in the understanding of interfacial crystallisation will also benefit nanoemulsion design. Due to their higher surface/volume ratio, oil droplets in nanoemulsion are more likely to be affected by interfacial effects and future work focussing on this issue is expected. The same additive may have a different effect depending on concentration, processing conditions and the presence of other additives [103]. In emulsified systems, it is very likely that the local concentration at the interface differs from the bulk concentration, as amphiphilic molecules below critical micelle concentration (CMC) are more likely to concentrate at the interface. Furthermore, emulsifiers are usually added in excess, therefore possibly interacting with growing fat crystals and affecting sintering/network formation. References [1] Alexa RI, Mounsey JS, O'Kennedy BT, Jacquier JC. Effect of κ-carrageenan on rheological properties, microstructure, texture and oxidative stability of water-in-oil spreads. LWT Food Sci Technol 2010;43:843–8. [2] McClements DJ. Crystals and crystallization in oil-in-water emulsions: implications for emulsion-based delivery systems. Adv Colloid Interface Sci 2012;174:1–30. [3] Rousseau D. Fat crystals and emulsion stability - a review. Food Res Int 2000;33:3–14. [4] Ghosh S, Rousseau D. Fat crystals and water-in-oil emulsion stability. Curr Opin Colloid Interface Sci 2011;16:421–31. [5] Coupland JN. Crystallization in emulsions. Curr Opin Colloid Interface Sci 2002;7:445–50. [6] Metin S, Hartel RW. Crystallization of fats and oils. In: Shahidi F, editor. Bailey's industrial oil and fat products. Hoboken: John Wiley & Sons, Inc.; 2005. p. 45–76. [7] Himawan C, MacNaughtan W, Farhat IA, Stapley AGF. Polymorphic occurrence and crystallization rates of tristearin/tripalmitin mixtures under non-isothermal conditions. Eur J Lipid Sci Technol 2007;109:49–60. [8] Marangoni AG, Acevedo N, Maleky F, Co E, Peyronel F, Mazzanti G, et al. Structure and functionality of edible fats. Soft Matter 2012;8:1275–300. [9] Rye GG, Litwinenko JW, Marangoni AG. Fat crystal networks. Bailey's industrial oil and fat products. John Wiley & Sons, Inc.; 2005 [10] Himawan C, Starov VM, Stapley AGF. Thermodynamic and kinetic aspects of fat crystallization. Adv Colloid Interface Sci 2006;122:3–33. [11] Acevedo NC, Peyronel F, Marangoni AG. Nanoscale structure intercrystalline interactions in fat crystal networks. Curr Opin Colloid Interface Sci 2011;16:374–83.

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10 [12] Acevedo NC, Marangoni AG. Characterization of the nanoscale in triacylglycerol crystal networks. Cryst Growth Des 2010;10:3327–33. [13] Acevedo NC, Marangoni AG. Toward nanoscale engineering of triacylglycerol crystal networks. Cryst Growth Des 2010;10:3334–9. [14] Sato K, Ueno S. Polymorphism in fats and oils. In: Shahidi F, editor. Bailey's industrial oil and fat products. Hoboken: John Wiley & Sons, Inc.; 2005. p. 77–120. [15] Sato K. Crystallization behaviour of fats and lipids — a review. Chem Eng Sci 2001;56:2255–65. [16] Trumbetas J, Fioriti J, Sims R. Applications of pulsed NMR to fatty emulsions. J Am Oil Chem Soc 1976;53:722–6. [17] van Boekel MA. Estimation of solid–liquid ratios in bulk fats and emulsions by pulsed nuclear magnetic resonance. J Am Oil Chem Soc 1981;58:768–72. [18] Boode K, Bisperink C, Walstra P. Destabilization of O/W emulsions containing fat crystals by temperature cycling. Colloids Surf 1991;61:55–74. [19] Fredrick E, Van de Walle D, Walstra P, Zijtveld J, Fischer S, Van Der Meeren P, et al. Isothermal crystallization behaviour of milk fat in bulk and emulsified state. Int Dairy J 2011;21:685–95. [20] Relkin P, Ait-Taleb A, Sourdet S, Fosseux P-Y. Thermal behavior of fat droplets as related to adsorbed milk proteins in complex food emulsions. A DSC study. J Am Oil Chem Soc 2003;80:741–6. [21] Higami M, Ueno S, Segawa T, Iwanami K, Sato K. Simultaneous synchrotron radiation X-ray diffraction - DSC analysis of melting and crystallization behavior of trilauroylglycerol in nanoparticles of oil-in-water emulsion. J Am Oil Chem Soc 2003;80:731–9. [22] Kalnin D, Schafer O, Amenitsch H, Ollivon M. Fat crystallization in emulsion: influence of emulsifier concentration on triacylglycerol crystal growth and polymorphism. Cryst Growth Des 2004;4:1283–93. [23] Relkin P, Yung J-M, Kalnin D, Ollivon M. Structural behaviour of lipid droplets in protein-stabilized nano-emulsions and stability of α-tocopherol. Food Biophys 2008;3:163–8. [24] Lopez C, Ollivon M. Crystallisation of triacylglycerols in nanoparticles. J Therm Anal Calorim 2009;98:29–37. [25] McClements DJ. Ultrasonic determination of depletion flocculation in oil-in-water emulsions containing a non-ionic surfactant. Colloids Surf A Physicochem Eng Asp 1994;90:25–35. [26] Kloek W, Walstra P, Vliet T. Nucleation kinetics of emulsified triglyceride mixtures. J Am Oil Chem Soc 2000;77:643–52. [27] Awad TS. Ultrasonic studies of the crystallization behavior of two palm fats O/W emulsions and its modification. Food Res Int 2004;37:579–86. [28] ten Grotenhuis E, Van Aken G, Van Malssen K, Schenk H. Polymorphism of milk fat studied by differential scanning calorimetry and real-time X-ray powder diffraction. J Am Oil Chem Soc 1999;76:1031–9. [29] Lopez C, Riaublanc A, Lesieur P, Bourgaux C, Keller G, Ollivon M. Definition of a model fat for crystallization-in-emulsion studies. J Am Oil Chem Soc 2001;78:1233–44. [30] MacNaughtan W, Farhat I, Himawan C, Starov V, Stapley AG. A differential scanning calorimetry study of the crystallization kinetics of tristearin-tripalmitin mixtures. J Am Oil Chem Soc 2006;83:1–9. [31] Höhne GWH. Fundamentals of differential scanning calorimetry and differential thermal analysis. In: Mathot VBF, editor. Calorimetry and thermal analysis of polymers. Munich, Vienna, New York: Hanser Publisher; 1997. p. 47–91. [32] Wunderlich B. Macromolecular physics, vol. 2 crystal nucleation, growth, annealing. New York: Academic Press; 1976. Kristall und Technik 1976;12:K11–2. [33] Avrami M. Kinetics of phase change. I General theory. J Chem Phys 1939;7:1103. [34] Turnbull D, Fisher J. Rate of nucleation in condensed systems. J Chem Phys 1949;17:71. [35] Foubert I, Fredrick E, Vereecken J, Sichien M, Dewettinck K. Stop-and-return DSC method to study fat crystallization. Thermochim Acta 2008;471:7–13. [36] Awad T, Sato K. Effects of hydrophobic emulsifier additives on crystallization behavior of palm mid fraction in oil-in-water emulsion. J Am Oil Chem Soc 2001;78:837–42. [37] Suryanarayana C, Norton MG. X-ray diffraction: a practical approach. Springer; 1998. [38] Müller-Buschbaum P. Grazing incidence small-angle X-ray scattering: an advanced scattering technique for the investigation of nanostructured polymer films. Anal Bioanal Chem 2003;376:7. [39] Singh MA, Groves MN. Depth profiling of polymer films with grazing-incidence small-angle X-ray scattering. Acta Crystallogr A 2009;65:190–201. [40] Tikhonov AM, Patel H, Garde S, Schlossman ML. Tail ordering due to headgroup hydrogen bonding interactions in surfactant monolayers at the water–oil interface. J Phys Chem B 2006;110:19093–6. [41] Levine JR, Cohen JB, Chung YW, Georgopoulos P. Grazing-incidence small-angle X-ray scattering: new tool for studying thin film growth. J Appl Crystallogr 1989;22:528–32. [42] Renaud G, Lazzari R, Leroy F. Probing surface and interface morphology with Grazing Incidence Small Angle X-Ray Scattering. Surf Sci Rep 2009;64:255–380. [43] Kalnin D, Garnaud G, Amenitsch H, Ollivon M. Monitoring fat crystallization in aerated food emulsions by combined DSC and time-resolved synchrotron X-ray diffraction. Food Res Int 2002;35:927–34. [44] Ollivon M, Keller G, Bourgaux C, Kalnin D, Villeneuve P, Lesieur P. DSC and high resolution X-ray diffraction coupling. J Therm Anal Calorim 2006;85:219–24. [45] Shinohara Y, Takamizawa T, Ueno S, Sato K, Kobayashi I, Nakajima M, et al. Microbeam X-ray diffraction analysis of interfacial heterogeneous nucleation of n-hexadecane inside oil-in-water emulsion droplets. Cryst Growth Des 2008;8:3123–6. [46] Wassell P, Okamura A, Young NWG, Bonwick G, Smith C, Sato K, et al. Synchrotron radiation macrobeam and microbeam X-ray diffraction studies of interfacial crystallization of fats in water-in-oil emulsions. Langmuir 2012;28:5539–47.

9

[47] Boode K, Walstra P. Partial coalescence in oil-in-water emulsions 1. Nature of the aggregation. Colloids Surf A Physicochem Eng Asp 1993;81:121–37. [48] Lopez C, Lesieur P, Keller G, Ollivon M. Thermal and structural behavior of milk fat - 1. Unstable species of cream. J Colloid Interface Sci 2000;229:62–71. [49] Ogden LG, Rosenthal AJ. Interactions between fat crystal networks and sodium caseinate at the sunflower oil–water interface. J Am Oil Chem Soc 1998;75:1841–7. [50] Ogden LG, Rosenthal AJ. Interactions between tristearin crystals and proteins at the oil–water interface. J Colloid Interface Sci 1997;191:38–47. [51] Binks BP. Particles as surfactants—similarities and differences. Curr Opin Colloid Interface Sci 2002;7:21–41. [52] Ghosh S, Rousseau D. Triacylglycerol Interfacial Crystallization and Shear Structuring in Water-in-Oil Emulsions. Crystal Growth & Design 2012;12:4944–54. [53] Ghosh S, Rousseau D. Freeze-thaw stability of water-in-oil emulsions. J Colloid Interface Sci 2009;339:91–102. [54] Frasch-Melnik S, Norton IT, Spyropoulos F. Fat-crystal stabilised w/o emulsions for controlled salt release. J Food Eng 2010;98:437–42. [55] Krog N, Larsson K. Crystallization at interfaces in food emulsions – a general phenomenon. Lipid/Fett 1992;94:55–7. [56] Garti N, Binyamin H, Aserin A. Stabilization of water-in-oil emulsions by submicrocrystalline alpha-form fat particles. J Am Oil Chem Soc 1998;75:1825–31. [57] Johansson D, Bergenståhl B, Lundgren E. Wetting of fat crystals by triglyceride oil and water. 1. The effect of additives. J Am Oil Chem Soc 1995;72:921–31. [58] Johansson D, Bergenståhl B. Lecithins in oil-continuous emulsions. Fat crystal wetting and interfacial tension. J Am Oil Chem Soc 1995;72:205–11. [59] Ghosh S, Rousseau D. Comparison of Pickering and network stabilization in waterin-oil emulsions. Langmuir 2011;27:6589–97. [60] Binks BP, Rocher A. Effects of temperature on water-in-oil emulsions stabilised solely by wax microparticles. J Colloid Interface Sci 2009;335:94–104. [61] Hodge SM, Rousseau D. Flocculation and coalescence in water-in-oil emulsions stabilized by paraffin wax crystals. Food Res Int 2003;36:695–702. [62] Street CB, Yarovoy Y, Wagner NJ, Vethamuthu MS, Hermanson KD, Ananthapadmanabhan KP. TDNMR characterization of a model crystallizing surfactant system. Colloids Surf A Physicochem Eng Asp 2012;406:13–23. [63] Campbell SD, Goff HD, Rousseau D. Comparison of crystallization properties of a palm stearin/canola oil blend and lard in bulk and emulsified form. Food Res Int 2002;35:935–44. [64] Ueno S, Hamada Y, Sato K. Controlling polymorphic crystallization of n-alkane crystals in emulsion droplets through interfacial heterogeneous nucleation. Cryst Growth Des 2003;3:935–9. [65] Katsuragi T, Kaneko N, Sato K. Effects of addition of hydrophobic sucrose fatty acid oligoesters on crystallization rates of n-hexadecane in oil-in-water emulsions. Colloids Surf B Biointerfaces 2001;20:229–37. [66] Kaneko N, Horie T, Ueno S, Yano J, Katsuragi T, Sato K. Impurity effects on crystallization rates of n-hexadecane in oil-in-water emulsions. J Cryst Growth 1999;197:263–70. [67] Awad T, Sato K. Acceleration of crystallisation of palm kernel oil in oil-in-water emulsion by hydrophobic emulsifier additives. Colloids Surf B Biointerfaces 2002;25:45–53. [68] Hindle S, Povey MJW, Smith K. Kinetics of crystallization in n-hexadecane and cocoa butter oil-in-water emulsions accounting for droplet collision-mediated nucleation. J Colloid Interface Sci 2000;232:370–80. [69] McClements DJ, Dungan SR, German JB, Simoneau C, Kinsella JE. Droplet size and emulsifier type affect crystallization and melting of hydrocarbon-in-water emulsions. J Food Sci 1993;58:1148–51. [70] Wright A, Marangoni A. Effect of DAG on milk fat TAG crystallization. J Am Oil Chem Soc 2002;79:395–402. [71] Chaleepa K, Szepes A, Ulrich J. Effect of additives on isothermal crystallization kinetics and physical characteristics of coconut oil. Chem Phys Lipids 2010;163:390–6. [72] Smith P, Furó I, Smith K, Cain F. The effect of partial acylglycerols on the exchange between liquid and solid tripalmitoylglycerol. J Am Oil Chem Soc 2007;84:325–9. [73] Gülseren İ, Coupland J. Surface melting in alkane emulsion droplets as affected by surfactant type. J Am Oil Chem Soc 2008;85:413–9. [74] Gülseren İ, Coupland J. The effect of emulsifier type and droplet size on phase transitions in emulsified even-numbered b i N nb/i N -alkanes. J Am Oil Chem Soc 2007;84:621–9. [75] McClements DJ, Dickinson E, Dungan SR, Kinsella JE, Ma JG, Povey MJW. Effect of emulsifier type on the crystallization kinetics of oil-in-water emulsions containing a mixture of solid and liquid droplets. J Colloid Interface Sci 1993;160:293–7. [76] Relkin P, Sourdet S, Fosseux PY. Fat crystallization in complex food emulsions: effects of adsorbed milk proteins and of a whipping process. J Therm Anal Calorim 2003;71:187–95. [77] Arima S, Ueno S, Ogawa A, Sato K. Scanning microbeam small-angle X-ray diffraction study of interfacial heterogeneous crystallization of fat crystals in oil-in-water emulsion droplets. Langmuir 2009;25:9777–84. [78] Sakamoto M, Ohba A, Kuriyama J, Maruo K, Ueno S, Sato K. Influences of fatty acid moiety and esterification of polyglycerol fatty acid esters on the crystallization of palm mid fraction in oil-in-water emulsion. Colloids Surf B Biointerfaces 2004;37:27–33. [79] Awad T, Hamada Y, Sato K. Effects of addition of diacylglycerols on fat crystallization in oil-in-water emulsion. Eur J Lipid Sci Technol 2001;103:735–41. [80] Mao L, O'Kennedy BT, Roos YH, Hannon JA, Miao S. Effect of monoglyceride selfassembled structure on emulsion properties and subsequent flavor release. Food Res Int 2012;48:233–40. [81] Ghosh S, Rousseau D. Triacylglycerol interfacial crystallization and shear structuring in water-in-oil emulsions. Cryst Growth Des 2012;12:4944–54.

10

M. Douaire et al. / Advances in Colloid and Interface Science 203 (2014) 1–10

[82] Tanaka L, Tanaka K, Yamato S, Ueno S, Sato K. Microbeam X-ray diffraction study of granular crystals formed in water-in-oil emulsion. Food Biophys 2009;4:331–9. [83] Davies E, Dickinson E, Bee R. Shear stability of sodium caseinate emulsions containing monoglyceride and triglyceride crystals. Food Hydrocoll 2000;14:145–53. [84] Smith K, Bhaggan K, Talbot G, van Malssen K. Crystallization of fats: influence of minor components and additives. J Am Oil Chem Soc 2011;88:1085–101. [85] Garti N, Aronhime J, Sarig S. The role of chain length and an emulsifier on the polymorphism of mixtures of triglycerides. J Am Oil Chem Soc 1989;66:1085–9. [86] Arima S, Ueji T, Ueno S, Ogawa A, Sato K. Retardation of crystallization-induced destabilization of PMF-in-water emulsion with emulsifier additives. Colloids Surf B Biointerfaces 2007;55:98–106. [87] Nik AM, Langmaid S, Wright AJ. Nonionic surfactant and interfacial structure impact crystallinity and stability of beta-carotene loaded lipid nanodispersions. J Agric Food Chem 2012;60:4126–35. [88] Bunjes H, Koch MHJ, Westesen K. Influence of emulsifiers on the crystallization of solid lipid nanoparticles. J Pharm Sci 2003;92:1509–20. [89] Helgason T, Awad T, Kristbergsson K, McClements DJ, Weiss J. Effect of surfactant surface coverage on formation of solid lipid nanoparticles (SLN). J Colloid Interface Sci 2009;334:75–81. [90] Jenning V, Mäder K, Gohla SH. Solid lipid nanoparticles (SLN) based on binary mixtures of liquid and solid lipids: a (1)H-NMR study. Int J Pharm 2000;205:15–21. [91] Shiota M, Iwasawa A, Kotera M, Konno M, Isogai T, Tanaka L. Effect of fatty acid composition of monoglycerides and shear on the polymorph behavior in waterin-palm oil-based blend. J Am Oil Chem Soc 2011;88:1103–11. [92] Aronhime J, Sarig S, Garti N. Mechanistic considerations of polymorphic transformations of tristearin in the presence of emulsifiers. J Am Oil Chem Soc 1987;64:529–33.

[93] Povey MJW, Hindle SA, Aarflot A, Hoiland H. Melting point depression of the surface layer in n-alkane emulsions and its implications for fat destabilization in ice cream. Cryst Growth Des 2005;6:297–301. [94] Montenegro R, Antonietti M, Mastai Y, Landfester K. Crystallization in miniemulsion droplets. J Phys Chem B 2003;107:5088–94. [95] Macierzanka A, Szeląg H, Moschakis T, Murray BS. Phase transitions and microstructure of emulsion systems prepared with acylglycerols/zinc stearate emulsifier. Langmuir 2006;22:2487–97. [96] Macierzanka A, Szeląg H. Microstructural behavior of water-in-oil emulsions stabilized by fatty acid esters of propylene glycol and zinc fatty acid salts. Colloids Surf A Physicochem Eng Asp 2006;281:125–37. [97] Johansson D, Bergenståhl B. Sintering of fat crystal networks in oil during postcrystallization processes. J Am Oil Chem Soc 1995;72:911–20. [98] Foubert I, Vanhoutte B, Dewettinck K. Temperature and concentration dependent effect of partial glycerides on milk fat crystallization. Eur J Lipid Sci Technol 2004;106:531–9. [99] Harada T, Yokomizo K. Demulsification of oil-in-water emulsion under freezing conditions: effect of crystal structure modifier. J Am Oil Chem Soc 2000;77:859–64. [100] Smith PR. The effects of phospholipids on crystallisation and crystal habit in triglycerides. Eur J Lipid Sci Technol 2000;102:122–7. [101] Bunjes H, Koch MHJ. Saturated phospholipids promote crystallization but slow down polymorphic transitions in triglyceride nanoparticles. J Control Release 2005;107:229–43. [102] Westesen K, Siekmann B. Investigation of the gel formation of phospholipidstabilized solid lipid nanoparticles. Int J Pharm 1997;151:35–45. [103] Rousseau D, Hodge SM, Nickerson MT, Paulson AT. Regulating the β′ → β polymorphic transition in food fats. J Am Oil Chem Soc 2005;82:7–12.