Crystals and crystallization in oil-in-water emulsions: Implications for emulsion-based delivery systems

Advances in Colloid and Interface Science 174 (2012) 1–30

Contents lists available at SciVerse ScienceDirect

Advances in Colloid and Interface Science

j ourna l homepage: www.e lsev ie r .com/ locate /c is

Crystals and crystallization in oil-in-water emulsions: Implications foremulsion-based delivery systems

David Julian McClements ⁎Department of Food Science, University of Massachusetts Amherst, Amherst, MA 01003, United States

⁎ Tel.: +1 413 545 1019; fax: +1 413 545 1262.E-mail address: [email protected].

0001-8686/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.cis.2012.03.002

a b s t r a c t
a r t i c l e i n f o
Available online 14 March 2012

Keywords:EmulsionNanoemulsionCrystallizationNucleationDissolutionEncapsulationPartitioningReleaseBioactivityNutraceuticalsFunctional foods

Many bioactive components intended for oral ingestion (pharmaceuticals and nutraceuticals) are hydropho-bic molecules with low water-solubilities and high melting points, which poses considerable challenges tothe formulation of oral delivery systems. Oil-in-water emulsions are often suitable vehicles for the encapsu-lation and delivery of this type of bioactive component. The bioactive component is usually dissolved in a car-rier lipid phase by either dilution and/or heating prior to homogenization, and then the carrier lipid andwater phases are homogenized to form an emulsion consisting of small oil droplets dispersed in water. Thesuccessful development of this kind of emulsion-based delivery system depends on a good understandingof the influence of crystals on the formation, stability, and properties of emulsions. This review article ad-dresses the physicochemical phenomena associated with the encapsulation, retention, crystallization, re-lease, and absorption of hydrophobic bioactive components within emulsions. This knowledge will beuseful for the rational formulation of effective emulsion-based delivery systems for oral delivery of crystallinehydrophobic bioactive components in the food, health care, and pharmaceutical industries.

© 2012 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.1. Emulsion-based delivery systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2. Encapsulation of crystalline bioactive components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2. Physicochemical aspects of solid–liquid transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.1. Dispersion of crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1.1. Melting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.1.2. Dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.1.3. Melting–dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.1.4. Crystal disruption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2. Supercooling and supersaturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.1. Supercooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.2.2. Supersaturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.3. Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.3.1. Nucleation within a supercooled melt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.3.2. Nucleation within a supersaturated solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.3.3. Nucleation mechanism: homogeneous versus heterogeneous nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.3.4. Nucleation promoters and inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.4. Crystal growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.5. Crystal morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112.6. Polymorphism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3. Nucleation and growth in emulsions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.1. Nucleation in emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3.1.1. Homogeneous nucleation within droplets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.1.2. Volume heterogeneous nucleation within droplets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.1.3. Mixed nucleation within droplets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

rights reserved.

http://dx.doi.org/10.1016/j.cis.2012.03.002

mailto:[email protected]

http://dx.doi.org/10.1016/j.cis.2012.03.002

http://www.sciencedirect.com/science/journal/00018686

2 D.J. McClements / Advances in Colloid and Interface Science 174 (2012) 1–30

3.1.4. Surface heterogeneous nucleation within droplets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.1.5. Nucleation within the aqueous phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3.2. Crystal growth and properties within emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.2.1. Crystal size and morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.2.2. Mass transport processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153.2.3. Polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153.2.4. Crystal location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3.3. Influence of crystals on emulsion stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174. Encapsulation of hydrophobic bioactive agents in emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4.1. Incorporation into the oil phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174.1.1. Melting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.1.2. Dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.1.3. Melting–dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

4.2. Encapsulation using high energy methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194.3. Encapsulation using low energy methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

5. The partitioning of bioactive substances in emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205.1. Maximum amount of solute that can be dissolved in oil–water system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215.2. Solute partitioning below the saturation level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215.3. Solute partitioning above the saturation level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225.4. Influence of carrier lipid properties on solute partitioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225.5. Influence of particle size on partitioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

6. Chemical stability of encapsulated bioactive components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237. Biological fate and bioavailability of encapsulated crystalline materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

7.1. Behavior of emulsion-based delivery systems within the GI tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237.1.1. Ingestion and digestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237.1.2. Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

7.2. Factors affecting biological fate of bioactive components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257.2.1. Transit time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257.2.2. Carrier oil digestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267.2.3. Precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267.2.4. Dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267.2.5. Solubilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267.2.6. Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

8. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

1. Introduction

This article focuses on the development of colloidal systems forthe delivery of hydrophobic bioactive components intended for oralconsumption, i.e., pharmaceuticals and nutraceuticals. Colloid deliv-ery systems consist of small particles dispersed within a carrier liquid,with the active component being encapsulated by the particle [1].Conventionally, a colloid is defined as having a characteristic dimen-sion between 1 and 1000 nm [2], but the term “colloidal deliverysystem” is often used more broadly to include other particulate sys-tems with similar structures and properties but somewhat largerparticle dimensions, e.g., diameters b10 μm. A variety of colloidalsystems have been developed for oral delivery applications in thepharmaceutical and food industries, including microemulsions, li-posomes, nanoemulsions, emulsions, multilayer emulsions, multi-ple emulsions, solid lipid nanoparticles, nanocrystal suspensions,and polymeric particles [3–8]. These delivery systems are designedto encapsulate, protect, and release hydrophobic bioactive compo-nents [9–15]. Many of these bioactive components are crystallineat ambient or body temperatures (Table 1), which causes chal-lenges to their effective utilization within commercial products.Crystalline materials are often difficult to incorporate into liquidproducts because of their tendency to aggregate and sediment,which causes problems with consistent dosing since some of thebioactive component may not be ingested. The presence of crystal-line materials may also adversely affect the appearance, texture,and mouthfeel of commercial products so that consumers are

unwilling to consume them, leading to poor compliance withdaily recommended intakes. Finally, the absorption of hydrophobicbioactive components from the gastrointestinal tract (GIT) may belimited when they are crystalline because this reduces their solu-bility and permeability [16–18]. The rational design of colloidal de-livery systems to encapsulate crystalline hydrophobic bioactivecomponents requires an understanding of the major factors that in-fluence their formation, stability, properties, and performance. Thisreview article focuses on the development of oil-in-water emulsionsas colloidal delivery systems for crystalline hydrophobic bioactivecomponents. It also discusses the potential impact of crystallizationof the carrier lipid matrix on the properties of colloidal deliverysystems.

1.1. Emulsion-based delivery systems

Oil-in-water emulsions consist of small oil droplets dispersed inan aqueous medium, with each droplet being coated by a thin layerof emulsifier molecules [4,19]. Unlike microemulsions, emulsionsare thermodynamically unstable colloidal dispersions, i.e., the free en-ergy of the separated oil and water phases is lower than that of theemulsion itself [4,6]. Consequently, they tend to breakdown overtime through processes such as flocculation, coalescence, gravitation-al separation, phase inversion, and Ostwald ripening [20]. The rate atwhich these processes occur depends on the physicochemical proper-ties of the two phases (e.g., polarity, density, and viscosity), the natureof the interfacial layer (e.g., thickness, charge, packing, chemistry),

Table 1Physicochemical properties of selected (mainly crystalline) hydrophobic components claimed to have health benefits when used as nutraceuticals (From SciFinder, AmericanChemical Society). The data is reported at pH 7 and 25 °C unless otherwise stated. Key: Tm is the melting temperature; ΔHf is the melting enthalpy, LogP is the logarithm of theoil–water partition coefficient; and CSW* is the water solubility.

Compound Molar mass Tm ΔHf LogP Molar volume CSW* pKa

(g mol−1) (°C) (J g−1) (cm3 mol−1) (g L−1)

Phenolic acidsCurcumin 368.4 183 120 3.07 288 0.052 8.1Capsaicin 305.4 64 3.20 293 0.14 9.9Vanillin 152.2 80 1.21 124 4.3 7.8

FlavonoidsCatechin 290.3 214 0.61 266 7.3 9.5Quercetin 302.2 310 1.99 168 1.9 6.3Genistein 270.2 300 3.11 175 0.7 6.5

StilbenoidsResveratrol 228.2 255 3.02 168 0.021 9.2

Cartenoidsβ-carotene 536.9 180 93 14.76 570 1.5×10−9 –

Lycopene 536.9 175 85 14.5 604 3.7×10−9 –

Zeaxanthin 568.9 205 10.9 564 8.0×10−5 14.6Astaxanthin 596.8 210 8.24 557 5.6×10−7 12.3Lutein 568.9 180 11.5 566 5.6×10−5 14.6

Phytosterolsβ-Sitosterol 414.7 138 46 10.5 424 8.7×10−7 15.0Campesterol 400.7 155 9.97 408 2.0×10−6 15.0Stigmasterol 412.7 150 63 10.07 418 1.6×10−6 15.0

Oil-soluble vitaminsα-tocopherol 430.7 3 10.96 463 5.2×10−5 11.4Tocopherol acetate 472.7 −27.5 10.69 502 2.0×10−6 –

MiscellaneousCoEnzyme-Q 863.3 49 19.12 888.5 5.9×10−8 –

3D.J. McClements / Advances in Colloid and Interface Science 174 (2012) 1–30

and particle characteristics (e.g., particle size, concentration andphysical state) [2,20]. Emulsions may be optically transparent or opa-que depending on the dimensions of the particles relative to thewavelength of light [21,22]. The rheological properties of emulsionsrange from low viscosity fluids, to highly viscous liquids, to solid-like materials depending on droplet concentration and interactions[23,24]. Consequently, it is possible to create emulsion-based deliverysystems with a wide range of optical, rheological, and functionalproperties by controlling their composition and microstructure.

1.2. Encapsulation of crystalline bioactive components

Oil-in-water emulsions are typically prepared by homogenizing alipid phase and an aqueous phase together in the presence of awater-soluble emulsifier [4,6]. This process leads to the formation ofemulsifier-coated oil droplets dispersed within an aqueous liquid. Acrystalline hydrophobic bioactive component may comprise the en-tire lipid phase or it may be dispersed within a carrier lipid prior toemulsion formation. The carrier lipid itself may be either liquid orsolid at the final application temperature [8]. Nevertheless, it is usual-ly important to ensure that the entire lipid phase (bioactive compo-nent+carrier lipid) remains liquid throughout the homogenizationprocess, which can be achieved using two different approaches:(i) the hydrophobic bioactive component is used below its saturationconcentration in a liquid carrier liquid; or, (ii) the lipid phase is heat-ed to melt any crystalline material before homogenizing. In addition,it is important to ensure that the delivery system remains stablethroughout its lifetime. The presence of crystalline material in anemulsion often promotes instability during storage [25,26], and so itis necessary to either avoid crystallization or to ensure that the emul-sion remains stable even if crystallization does occur. Finally, it is im-portant to ensure that any bioactive component encapsulated withinan emulsion-based delivery system is actually released within the

gastrointestinal tract (GIT) and absorbed by the body otherwise itsbioactivity will be reduced [16,17]. Crystalline hydrophobic compo-nents usually have to dissolve and be incorporated into mixed mi-celles or other colloidal structures before they can be absorbed bythe human body. Crystals may be present in the original delivery sys-tem or they may form within the gastrointestinal tract due to precip-itation [27]. Hence it is important to understand the factors thatinfluence the formation and dissolution of crystals within the GIT inorder to optimize the bioavailability of high melting point hydropho-bic components. Finally, it may be possible for non-dissolved crystalsto be directly absorbed by epithelium cells if they are sufficientlysmall, which could influence their bioactivity [8].

This review article focuses on the physiochemical phenomena thatneed to be considered when designing emulsion-based delivery sys-tems that contain crystalline hydrophobic components. There are anumber of different kinds of delivery system that may be used to en-capsulate bioactive crystalline components (Fig. 1): (i) a crystal hy-drophobic bioactive component may be dissolved in either the oiland/or water phases; (ii) the carrier lipid surrounding a bioactivecomponent within a colloidal particle may be either fully or partiallycrystalline (as in solid lipid nanoparticles or nanostructured lipid car-riers); (iii) the bioactive component may be dispersed in the aqueouscontinuous phase entirely as crystals (as in nanocrystal suspensions).This article begins by discussing solid–liquid phase transitions inmelts and solutions, since this impacts the formation, loading capaci-ty, long-term stability, functional performance, and biological fate ofemulsion-based delivery systems. The methods available for creatingemulsion-based delivery systems for encapsulating crystalline bioac-tive components are then briefly reviewed, and some of the majorfactors that influence their stability and physical properties arehighlighted. Finally, how the bioavailability of crystalline hydropho-bic bioactive components may be influenced by the behavior of emul-sions within the GI tract is considered.

NanocrystalSuspension

NanostructuredLipid Carrier

Solid Lipid Nanoparticle

LiquidMatrix

BioactiveCrystals

DisorderedCrystalline Matrix

OrderedCrystalline Matrix

Emulsion(C < CS)

Emulsion(C > CS)

DissolvedBioactive

Fig. 1. Examples of different kinds of emulsion-based delivery systems containing crystalline bioactive components. Either the bioactive component or the surrounding carrier lipidmay be crystalline.

GSL

Liquid

G*

Solid

T < Tm

Activation Energy

Disordered

Ordered

NucleiFormation

Δ

Δ

Fig. 2. Free energy changes (ΔG) associated with the solid–liquid phase transition.Below the melting temperature the difference in free energy between the solid and liq-uid phases (ΔGSL) is negative thereby favoring the solid state, but there may be an ac-tivation energy (ΔG*) associated with nuclei formation that must be overcome beforethe transition will occur.


2. Physicochemical aspects of solid–liquid transitions

The development of effective emulsion-based delivery systemscontaining crystalline bioactive components requires an understand-ing of crystal formation and characteristics within multiphasic sys-tems. In this section, an overview of the major factors influencingthe phase behavior and physicochemical properties of hydrophobiccrystals is given.

2.1. Dispersion of crystals

The preparation of oil-in-water emulsions usually requires thatthe oil phase remain liquid throughout the homogenization process.We therefore begin by considering the factors that influence the con-version of crystalline bioactive components into a liquid form suitablefor homogenization. Two different approaches can be used to ensurethat crystalline bioactive components are liquid during homogeniza-tion: melting and dissolution. It is possible to produce suspensionscontaining small crystals by passing large crystals suspended in anaqueous emulsifier solution through a high pressure homogenizerto disrupt them (i.e., nanocrystal suspensions) [8,28]. Technically,nanocrystal suspensions are not emulsion-based delivery systems,but they do have some similar features (emulsifier-coated lipid parti-cles dispersed in water) and so they are briefly considered here.

2.1.1. MeltingIn principle, an emulsion-based delivery system can be fabricated

from a crystalline bioactive component by heating it above its meltingpoint (Tm) to form a liquid lipid phase, and then homogenizing thislipid phase with an aqueous phase containing a water-soluble emul-sifier at a sufficiently high temperature (T>Tm). The resulting emul-sion could then be cooled below the melting point of the bioactivecomponent to promote a liquid–solid phase transition and form a sus-pension of nano-sized or micro-sized crystals dispersed in water. Inpractice, this approach is unsuitable for many hydrophobic bioactivecomponents because their melting points are too high (>200 °C) forcommercial homogenizers (Table 1), or because they (or other com-ponents in the system) are chemically unstable at elevated tempera-tures (e.g., carotenoids rapidly degrade at high temperatures).

Below the melting temperature, the most thermodynamically sta-ble form of a material is the solid form because the free energy of thesolid state is lower than that of the liquid state (Fig. 2). Nevertheless,there are usually kinetic energy barriers (“activation energies”) thatprevent a system from attaining its most thermodynamically stablestate. Kinetic effects are particularly important in determining therate and extent of crystal formation from melts and supersaturatedsolutions since there is an activation energy associated with nuclei

formation, which is a pre-requisite for crystal formation (seeSection 2.3).

2.1.2. DissolutionCrystalline hydrophobic bioactive components are often dissolved in

a solvent to ensure that the lipid phase remains liquid during homoge-nization. This solvent may be a carrier lipid (such as triacylglycerol oil)or an organic solvent (such as an alcohol or hydrocarbon). Some ofthe crystalline material may also dissolve in the aqueous phase eitherbefore or after emulsion formation depending on the equilibrium parti-tion coefficient and water-solubility (Section 5). It is therefore impor-tant to understand the major factors that influence the dissolution ofcrystals in solvents. From a thermodynamic point of view, a crystallinematerial has a finite solubility in a given solvent (CS⁎)—at equilibriumthe material is fully dissolved below this level, but forms crystalsabove this level. Knowledge of the saturation concentration of a solutein the various phases that a delivery system is fabricated from (andcomes into contact with) is therefore crucial for designing the system


to perform properly, i.e., to avoid crystal formation during storage orafter ingestion. The saturation concentration depends on the nature ofthe bioactive component, the nature of the solvent, and environmentalconditions (e.g., pH, ionic strength, and temperature). The molecularproperties of solutes can be related to their saturation concentrationsin various solvents using various empirical, theoretical, and computa-tional approaches [30–32]. At present no single approach can accuratelypredict the solubility of all solutes in all solvents, but many of themethods do produce fairly good estimations for small solutes. An ex-ample of a relatively simple method that can be used to relate thewater-solubility of solutes to easily measurable properties is theGeneral Solubility Equation (GSE): log(SW)=0.5−0.01 (Tm−25)−log(KOW), where SW is the molar water-solubility, Tm is the meltingpoint, and KOW is the oil–water partition coefficient of the solute[33,34]. This model indicates that the water-solubility of a substanceshould decrease as its melting point and KOW increase.

In practice, it is often possible to dissolve a greater amount of acrystalline material into a solvent than CS⁎ due to supersaturation. Asolution becomes supersaturated because of the activation energythat must be overcome before solute molecules in solution cancome together and assemble into the nuclei that must be formed be-fore crystals can grow (see Section 2.3). An appreciation of saturationand supersaturation phenomenon can be obtained by examining thenature of the solubility curves typically obtained experimentally. Inthese experiments, a series of samples is prepared by dispersing in-creasing amounts of crystalline solute into a particular solvent [36].At relatively low solute concentrations, the solute fully dissolves inthe solvent and the samples appear transparent, but at higher con-centrations some of the crystals remain undissolved and the samplesappear turbid or contain sediment. Any crystals present in a sampleare then removed (e.g. by filtration or centrifugation) and the dis-solved solute concentration in the supernatant is measured. A repre-sentative plot of the dissolved solute concentration (CD) versus thetotal solute concentration added (CT) is shown in Fig. 3. With increas-ing solute concentration, there is initially a linear increase in CD withCT because all the solute dissolves, but once a certain concentration isexceeded there is a sharp decrease in CD, after which CD remains rel-atively constant. The plateau value of CD observed at relatively highsolute concentrations is equivalent to the saturation solute concen-tration (CS⁎). The fact that the dissolved solute concentration can ex-ceed the saturation concentration is due to supersaturation. Thedegree of supersaturation in a system depends on solute type, solventtype, and environmental conditions such as temperature, mechanicalforces, and time.

Fig. 3. A representative plot from a typical solubility experiment showing the dissolvedsolute concentration (CD) versus the total solute concentration added (CT). At relativelylow concentrations the solute dissolves, but above a certain concentration crystals formin the sample, and the solute concentration is at the saturation level (CS*).

When performing such experiments it is important to take into ac-count that the rate of crystal dissolution may vary considerablydepending on the system. The rate at which crystals dissolve in a liq-uid depends on the nature of the crystals (such as surface area, crystalstructure), the nature of the solvent (such as polarity and molecularweight), and environmental conditions (such as temperature, sonica-tion, and stirring speed). The dissolution of crystals can vary from sec-onds to days depending on their sizes and solubilities (Table 2). It istherefore important to ensure that a system has come to equilibriumbefore carrying out measurements of the saturation concentration.This can be achieved by measuring the change in the dissolved soluteconcentration over time to determine the dissolution kinetics, andestablishing when saturation has been reached [36]. It is also impor-tant to establish the kinetics of crystal dissolution when designingpreparation methods for fabricating emulsion-based delivery systemscontaining solid bioactive agents. For example, we have found thathydrophobic bioactive components may form crystals in emulsion-based delivery systems during long-term storage, which sedimentto the bottom of the system, which is obviously undesirable for com-mercial applications [37].

The dissolution process can be considered to consist of two steps:(i) solvation and detachment of the solute molecules from the crystal;(ii) movement of the solute molecules away from the interfacethrough the surrounding solution [38]. It is usually assumed thatthe mass transport of solute molecules away from the interface isthe rate-limiting step, which means that the process can be modeledusing Fick's first law of diffusion. Under this assumption, the dissolu-tion process can be modeled using the Noyes–Whitney equation [38]:

dMdt

¼ A� Dh

CS⁎−CBð Þ ð1Þ

Where M is the mass of solute leaving the particle, t is time, A isthe surface area of the dissolving substance, D is the diffusion coeffi-cient of the solute in the surrounding liquid, h is the thickness ofthe diffusion layer, CS⁎ is the equilibrium solubility, and CB is the con-centration of solute in the surrounding solution. This equation indi-cates that a particle will continue to dissolve until the concentrationof solute in the surrounding solution reaches the solubility limit(CS⁎=CB). It should be noted that A and h both depend on time,and therefore this equation is difficult to solve for complex particlegeometries. However, a simple expression can be derived for thechange in the amount of material present in the solid state overtime for spherical particles [38]:

M1=3 ¼ M1=30 1−DCS⁎

hr0ρ

� �ð2Þ

Table 2Calculated dissolution times for spherical solid particles with different dimensions anddifferent equilibrium solubilities. Values were calculated using Eq. (3) assuming thatthe particles consisted of a model hydrophobic bioactive component: β-carotene,where ρ=1000 kg/m−3 and D=2×10−9. The saturation concentrations (CS) aregiven in weight percent.

Radius(μm)

Dissolution time (s)

CS=10−8 CS=10−6 CS=10−4 CS=10−2

0.01 2.5×102 2.5×100 2.5×10−2 2.5×10−4

0.1 2.5×104 2.5×102 2.5×100 2.5×10−2

1 2.5×106 2.5×104 2.5×102 2.5×100

10 2.5×108 2.5×106 2.5×104 2.5×102

100 2.5×1010 2.5×108 2.5×106 2.5×104

1000 2.5×1012 2.5×1010 2.5×108 2.5×106

Fig. 4. Plot of the change in solubility of a solid substance in a carrier oil as the temperatureis varied. The calculationswere carried out forβ-carotene dissolved in a triacylglycerol oil :ΔHf=76 J/g; Tm=181 °C; MS=537 g/mol; MO=800 g/mol. The predictions were madeusing Eqs. (4) and (5) described in the manuscript.


HereM is the mass of material remaining in the particles at time t,M0 is the initial mass of material in the particles, ρ is the density of thematerial in the particle, and r0 is the initial radius of the particle. Forsmall particles (rb25 μm), h can be taken to be approximately equalto the radius of the particle (i.e., h=r0). The following simple expres-sion has been derived to estimate the time required for small spheri-cal particles to dissolve in a well agitated aqueous solution [38]:

τ ¼ ρr202DCS⁎

ð3Þ

This equation indicates that the particles should dissolve morerapidly as the radius of the particles decreases, the diffusion coeffi-cient of the solute molecules in the surrounding medium increases(i.e., its viscosity decreases), and the equilibrium solubility of the sol-ute in the surrounding medium increases. As mentioned earlier, theabove equations have been derived for spheres, whereas most crys-talline materials have much more complex morphologies. Neverthe-less, these equations do provide some useful insights into the majorfactors that influence the dissolution process. Calculated dissolutiontimes for solid particles dispersed in water at ambient temperatureare shown in Table 2. In practice, the dissolution times may be consid-erably longer than these values if the rate-limiting step is detachmentof the solute molecules from the solid surface.

2.1.3. Melting–dissolutionA combination of dissolution and melting is often used to produce

a liquid lipid phase from a crystalline bioactive component and sol-vent (such as a carrier lipid). The solubility of a crystalline componentin a solvent can be predicted assuming they have widely differingmelting points (>20 °C) and form an ideal mixture:

x ¼ expΔHf

R1Tm

� 1T

� �� ð4Þ

Here x is the solubility (mole fraction) of the higher melting pointcomponent in the lower melting point component, T is the tempera-ture, Tm is the melting point, R is the gas constant, and ΔHf is themolar heat of fusion [26]. Above the melting point x=1, i.e., thehigher melting compound is completely liquid and miscible withthe lower melting component. The mole fraction of a solute dissolvedin the oil phase at saturation can be converted into a mass fraction(ΦS) if the molar masses of the two components are known:

ΦS ¼ 1þ MS

MO

1x−1

� �� −1ð5Þ

HereMS andMO are the molar masses of the solute and the carrieroil, respectively. This equation can be used to calculate the change insolubility of a crystalline material in carrier oil with temperature. Asthe temperature increases, the solubility of the crystalline materialin the carrier oil increases (Fig. 4). Consequently, it is possible to cre-ate a liquid lipid phase that contains a relatively high concentration ofsolute at elevated temperatures. An advantage of this method is that ahomogenization temperature that is considerably lower than that ofthe pure bioactive substance can be used to carry out the homogeni-zation. Nevertheless, some of the solute molecules may crystallizewhen the solution is cooled to a lower temperature, since the soluteconcentration may then exceed the saturation level at that tempera-ture. The rate of this process will depend on kinetic factors, such asnucleation and crystal growth rates (Section 2.3). In practice, theabove equations only have limited application because most real sys-tems are non-ideal mixtures, so that their solubility behavior dependson the precise nature of the molecular interactions involved [39]. Theequation for predicting the temperature dependence of a solute in asolvent (Eq. (4)) can be modified to take into account non-ideal

mixtures by dividing the right hand side by the activity coefficient(γi). A number of mathematical models have been developed to pre-dict the activity coefficients of solutes in solvents based on their mo-lecular characteristics, which have been reviewed elsewhere [31].

2.1.4. Crystal disruptionNanocrystal suspensions can be produced by passing micrometer-

sized crystals through a mechanical device that breaks them down,e.g., high pressure homogenizer, microfluidizer, milling devices[8,40]. These kinds of mechanical devices are finding increasing utili-zation in the pharmaceutical and food industry for the delivery of hy-drophobic crystalline bioactive components [8]. In this type ofdelivery system, it is not necessary to dissolve the crystals prior to ho-mogenization. Instead, existing crystals are broken down to smallerdimensions by the application of intense disruptive forces. Normally,the initially large crystals are dispersed in an aqueous phase contain-ing an emulsifier that adsorbs to the crystal surfaces and preventsthem from aggregating after mechanical disruption. In the authorsopinion there may be some advantages in incorporating the nano-crystals formed by this approach into emulsions instead of water.The presence of digestible oil often facilitates the absorption of bioac-tive components in the small intestine since it increases the amountof mixed micelles available to solubilize and transport hydrophobiccomponents [16,17].

2.2. Supercooling and supersaturation

As mentioned earlier, there are certain kinetic factors (energy bar-riers) that can prevent crystals from forming even under circum-stances where crystallization is thermodynamically favorable(Section 2.1). A melt has to be cooled below its melting point beforecrystals will form (“supercooling”), whereas the solute concentrationin a solution has to exceed the saturation level before crystals willform (“supersaturation”). A supercooled or supersaturated solutionmay persist for a considerable time before any crystallization is ob-served because of the activation energy that must be overcome beforethe liquid–solid phase transition can occur (Fig. 2). If the magnitudeof this activation energy is sufficiently high compared to the thermalenergy of the system, then crystallization does not occur on a reason-able timescale, even though the transition is thermodynamically fa-vorable [35]. The system is then said to exist in a metastable state.The height of the activation energy depends on the ability of crystalnuclei to be formed that are stable enough to grow into crystals(see Section 2.3).

image of Fig.�4


2.2.1. SupercoolingThe degree of supercooling of a liquid is usually defined as ΔT=

T−Tm, where T is the temperature and Tm is the thermodynamicmelting point. The value of ΔT at which the formation of crystalsis first observed depends on the nature of the lipid phase, the pres-ence of any contaminating materials, the cooling rate, the micro-structure of the lipid phase (e.g., bulk versus emulsified), and theapplication of external forces such as high pressure, shear, or son-ication [41]. Pure lipid phases containing no impurities can typical-ly be supercooled by more than 10–15 °C before any crystallizationis observed [42–45]. The phenomenon of supercooling is particu-larly important in emulsified oils, since each oil droplet has a lowprobability of containing any impurities and so nucleation oftenproceeds by a homogeneous nucleation (Section 3.1).

2.2.2. SupersaturationThe degree of supersaturation of a solution is usually defined as

S=CB/CS*, where CB is the total solute concentration and CS* is thesaturation concentration of the solute in the solvent. The degree ofsupersaturation that can be achieved before any crystals will formon a reasonable timescale depends on storage temperature, the na-ture of the solute and solvent, the presence of any contaminating ma-terials, and the application of external forces [36,46].

2.3. Nucleation

Crystals can only grow after stable nuclei have been formed in aliquid [36,46–48]. These nuclei are believed to consist of small tran-siently ordered molecular clusters that form when a number of mol-ecules collide and become associated with each other (Fig. 5).Nuclei can either dissociate or grow into crystals. A brief descriptionof the major factors that influence the formation of stable nuclei with-in supercooled melts and supersaturated solutions is presented in thissection.

2.3.1. Nucleation within a supercooled meltThe reason that crystals do not form immediately within a super-

cooled melt (even though one is below the melting point) can be

Classical NucleatioMonomers cluster to fo

Two-Step Nucleat

Monomers cluster to form condensed phase

M

Monomers in solution

Monomers

Fig. 5. Schematic diagram of the classical nucleation theory and two-step nucleation theory tMyerson (2006).

explained in terms of the free energy changes associated with nucleiformation [26,41,46]. Below the melting point, the bulk crystallinestate is thermodynamically more favorable than the bulk liquidstate, and so there is a decrease in free energy when some of the mol-ecules in the melt cluster together to form a nucleus. This negative(favorable) free energy (ΔGV) change is proportional to the volumeof the nucleus formed. On the other hand, the formation of a nucleusleads to the creation of a new interface between the solid phase andthe liquid phase which requires an input of free energy to overcomethe interfacial tension. This positive (unfavorable) free energy (ΔGS)change is proportional to the surface area of the nucleus formed.The total free energy change associated with the formation of a nucle-us is therefore a combination of a volume and a surface term [49,50]:

ΔG ¼ ΔGV þ ΔGS ¼43πr3

ΔHfΔTTm

þ 4πr2γSL ð6Þ

where r is the radius of the nuclei (assuming it is spherical), ΔHf is theenthalpy change per unit volume associated with the liquid–solidtransition (which is negative) and γSL is the solid–liquid interfacialtension (which is positive). The volume contribution becomes in-creasingly negative as the size of the nuclei increases, whereas thesurface contribution becomes increasingly positive (Fig. 6). The sur-face contribution dominates for small nuclei, while the volume termdominates for large nuclei. The overall free energy has a maximumvalue at a critical nucleus radius (r⁎):

r⁎ ¼ −2γSLTm

ΔHfΔTð7Þ

If a nucleus is formedwith a radius below this critical value (rb r⁎) itwill tend to dissociate so as to reduce the free energy of the system. Onthe other hand, if a nucleus is formed with a radius above this criticalvalue (r>r⁎) it will tend to grow into a crystal to reduce the free energyof the system. This equation indicates that the critical nuclei size re-quired for crystal growth decreases as the degree of supercooling in-creases (ΔT), which accounts for the increase in nucleation rate withdecreasing temperature that is typically observed experimentally

Crystals

n Theoryrm nuclei

ion Theory Monomers add to nuclei to form crystals

onomers rearrange within condensed phase

hat have been developed to model nucleation and crystal growth. Adapted from Lee and

image of Fig.�5

G*

GS

GV

r*

G

r

G

Δ

Δ Δ

Δ

Δ

Fig. 6. Schematic diagram of the free energy changes associated with the formation ofnuclei with different radii (r) in a supercooled melt. The overall free energy (ΔG) de-pends on a volume contribution (ΔGV) that favors nuclei formation and a surface con-tribution (ΔGS) that opposes nuclei formation. There is a free energy maximum (ΔG*)and critical radius (r*) associated with nuclei formation. Adapted from McClements(2005).

T*

Supercooling, ΔT

J

Inhibition due to retarded diffusion

Δ

Fig. 7. Schematic diagram of the influence of temperature (supercooling) on the nucle-ation rate (J) of a substance. The full line is the prediction without considering molec-ular diffusion effects, whereas the broken line takes into account reduced moleculardiffusion at low temperatures. Nucleation occurs at a relatively rapid rate when a crit-ical degree of supercooling (ΔT*) is exceeded.


below themelting point. The free energy associated with the formationof a nucleus of the critical size (ΔG⁎) is calculated by replacing r inEq. (6) with the critical radius given in Eq. (7) [26].

ΔG⁎ ¼ 16πγ3SLT

2m

3ΔH2f ΔT

2 ¼ 43πr⁎2γSL ð8Þ

The rate at which nuclei form in the melt is related to the activa-tion energy (ΔG⁎) that must be overcome before stable nuclei canbe formed is usually described by classical nucleation theory (CNT).

J ¼ J0 exp −ΔG⁎kbT

� �ð9Þ

The nucleation rate (J) is the number of stable nuclei formed persecond per unit volume of material, J0 is a pre-exponential factor, kBis Boltzmann's constant, and T is the absolute temperature. Expres-sions for the pre-exponential factor have been calculated based onthe assumption that there is one or more rate limiting steps (activa-tion energies) associated with the movement of molecules to the sur-face of the nuclei and with their incorporation into the nuclei [26]:

J0 ¼ N0kBTh

exp −ΔG⁎

Diff

kBT

" #ð10Þ

Here N0 is the number of molecules per unit volume that can un-dergo the phase transition, h is Plank's constant, and ΔGdiff⁎ is the ac-tivation energy associated with the diffusion of the hydrophobicmolecules to the solid–liquid interface [51]. The value of ΔGdiff⁎ de-pends on the viscosity of the liquid surrounding the nuclei, and there-fore tends to increase as the temperature decreases since then thehydrophobic molecules move more slowly to the interface.

The dependence of the nucleation rate predicted on the degree ofsupercooling (ΔT) is shown schematically in Fig. 7 in the absence andpresence of a diffusional barrier. The formation of stable nuclei is neg-ligibly slow at temperatures just below the melting point, but in-creases dramatically when the liquid is supercooled below a certaintemperature, T⁎. In the presence of a diffusional barrier, the nucle-ation rate increases with supercooling down to a certain temperature,

but then decreases upon further cooling because the increase in vis-cosity of the liquid phase that occurs as the temperature is decreasedslows down the diffusion of hydrophobic molecules towards theliquid-nucleus interface, i.e., increases ΔGdiff⁎ [41]. Consequently,there is usually a maximum in the nucleation rate at a particular tem-perature. The nucleation rates predicted by the classical nucleationtheory (CNT) may be orders of magnitudes different from experimen-tally measured ones, although CNT does provide valuable insightsinto the major factors influencing nucleation [26,41,52].

2.3.2. Nucleation within a supersaturated solutionClassical nucleation theory can also be used to describe nucleation

in supersaturated solutions. In this case, nuclei form within a solventthat has a different molecular structure to the solute, rather than in aliquid phase with the samemolecular structure as the nucleating sub-stance [46,53]. An expression for the critical radius for the formationof stable nuclei in the case of supersaturation can again be derivedfrom an analysis of the associated free energy changes:

r⁎ ¼ 2γSL Vm=NAð ÞkBT ln Sð Þ ð11Þ

The free energy associated with the formation of nuclei with thiscritical radius in a supersaturated solution is then given by:

ΔG⁎ ¼ 16πγ3SL Vm=NAð Þ2

3 kBT ln Sð Þð Þ2 ð12Þ

Here γSL is the solid–liquid interfacial tension, Vm is the molar vol-ume of the bioactive component, NA is Avogadro's number, and S isthe degree of supersaturation (S=CB/CS⁎). The rate at which nucle-ation proceeds is related to the activation energy ΔG⁎ that must beovercome before stable nuclei are formed:

J ¼ J0 exp −ΔG⁎kbT

� �ð13Þ

In this case, the pre-exponential factor has been given as [54]:

J0 ¼ βDS NACBð Þ2 kBTγSL

� �1=2ln Sð Þ ð14Þ


Where DS is the diffusion coefficient of the solute moleculesthrough the solvent phase, and β is a parameter that depends onwhether homogeneous or heterogeneous nucleation occurs [54].

It is informative to examine the practical implications of theseequations for the formation of nuclei in supersaturated solutions.First, the nucleation rate increases as the degree of supersaturation(S) in the system increases, since S>1, therefore ΔG⁎ decreases as Sincreases. Second, the nucleation rate has a strong dependence onthe interfacial tension, since ΔG⁎ is proportional to γSL cubed. The in-terfacial tension depends on the nature of the solvent that the solutemolecules are dispersed within (e.g., oil versus water), which mayhave important implications for situations where nucleation occursin multiphase systems such as oil-in-water emulsions. As the interfa-cial tension increases, the energy barrier for nuclei formation in-creases, which decreases the nucleation rate. Consequently, onewould expect nucleation of a hydrophobic component to occurmore rapidly in an oil phase (lower γSL) than in a water phase (higherγSL). Third, the nucleation rate increases as the diffusion coefficient ofsolute molecules in the surrounding solvent increases, and hence onewould expect faster nucleation in lower viscosity solvents. This phe-nomenon may also have important implications for situationswhere nucleation occurs in emulsions. One would expect nucleationof a hydrophobic component to occur more slowly in an oil phase(higher η) than in a water phase (lower η), which may counter-balance the interfacial tension effect mentioned earlier.

Despite the classical nucleation theory (CNT) being widely used todescribe nucleation from melts and solutions it is based on a numberof assumptions that limit its accuracy [55–58]. For example, it as-sumes: (i) the nuclei are spherical and have the same physicochemi-cal properties as bulk crystals; (ii) the interfacial tension does notdepend on nuclei size and has the same value as for a macroscopiccrystal–liquid interface; (iii) clusters are only formed by progressiveaddition of individual monomers (i.e., cluster–cluster fusion/fissiondoes not occur); (iv) the nucleation rate is constant over time;(v) the solute concentration in the liquid remains constant (at thesaturation level). Therefore, the CNT only usually provides qualitativeinsights into the major factors influencing nucleation rather thanquantitative predictions of nucleation rates. Indeed, measured nucle-ation rates are often orders of magnitudes different from those pre-dicted by CNT.

More recently a two-step model has been proposed as giving amore realistic description of nucleation and growth processes[55,57,58]. In this model, a cluster of solute molecules forms first,and then these solute molecules undergo a molecular rearrangementto form a more ordered structure that comprises the nuclei, whichthen grows into a crystal (Fig. 5). Mathematical models and computersimulations have recently been developed to describe this two-stepprocess, which often give better descriptions of experimental datathan CNT [29,57,58]. This model was initially developed for applica-tion to crystallization of proteins from solution, but it may also haveapplicability to other crystallizing substances.

The interfacial tension at the crystal–water boundary is an impor-tant parameter in the equations describing nucleation in aqueous so-lutions. It is often difficult to measure this interfacial tension usingconventional experimental methods, due to the difficulty in produc-ing large crystals with smooth surfaces. Recently it has been proposedthat the following equation can be used to predict the interfacial ten-sion at the crystal–water interface based on knowledge of the molec-ular and physicochemical characteristics of the crystal [59]:

γSL ¼ − 0:33kBTVm=NAð Þ2=3 ln

CS⁎

55:6

� �þ 5

� �ð15Þ

This equation assumes that the crystal–water interface is clean,i.e., there are no other substances absorbed to it. In practice, theremay often be surface-active substances in the aqueous phase that

can adsorb to the crystal–water interface and reduce the interfacialtension, such as surfactants or polymers. These substances wouldtherefore be expected to promote nucleation, since they wouldlower the activation energy for nuclei formation. On the other hand,they may inhibit crystal growth by preventing additional solute mol-ecules from adsorbing to and being incorporated into the growingcrystal surface. Nevertheless, the above equation can be used to givean approximate value for the solid–liquid interfacial tension formany substances once the equilibrium solubility and molar volumeof the solute are known.

2.3.3. Nucleation mechanism: homogeneous versus heterogeneousnucleation

The type of nucleation described above occurs when there are noimpurities present in the melt or solution, and is usually referred toas homogeneous nucleation [41]. If the liquid phase is in contact withimpurities, such as the surfaces of dust particles, crystals, oil droplets,air bubbles, reverse micelles, or the container, then nucleation canoccur at a lower degree of supercooling or supersaturation thanexpected for a pure system [26]. Nucleation due to the presence ofthese impurities is referred to as heterogeneous nucleation, and canbe divided into two types: primary and secondary [41,46]. Primaryheterogeneous nucleation occurs when the impurities have a differentchemical structure to that of the solute. Secondary heterogeneousnucleation occurs when the impurities are seed crystals with thesame chemical structure as the solute.

Heterogeneous nucleation occurs when the impurities provide asurface where the formation of stable nuclei is more thermodynami-cally favorable (ΔG⁎ is less) than in pure liquid [46]. As a result thedegree of supercooling or supersaturation required to initiate nucle-ation is reduced. On the other hand, certain types of impurities are ca-pable of decreasing the nucleation rate because they are incorporatedinto the surface of the growing nuclei and prevent any further mole-cules being incorporated [26,41]. Whether an impurity acts as a cata-lyst or an inhibitor of nucleation depends on its molecular structureand interactions with the nuclei (Section 2.3.4).

A number of mathematical expressions have been derived to pre-dict the heterogeneous nucleation rate. In general, for an impurity topromote nucleation it must lower the free energy for the formation ofnuclei in the system below that for spontaneous nucleation formationin the bulk. The tendency for nuclei to form on the surface of an im-purity (I) is mainly determined by the contact angle, which is gov-erned by the interfacial tensions at the solid-impurity, liquid-impurity and solid–liquid boundaries, where the solid (S) representsthe nuclei and the liquid (L) represents the surrounding liquid [50]:

cosθ ¼ γSI−γLI

γSLð16Þ

The same mathematical model can be used to describe heteroge-neous nucleation as was used to describe homogeneous nucleation(Sections 2.3.1 and 2.3.2), but with the interfacial tension (γ)replaced by an effective interfacial tension (γeff):

γSL;ef f ¼ γSL

ffiffiffiffiffiffiffiffiffiffiψ θð Þ3

pð17Þ

Where

ψ θð Þ ¼ 14

2þ cosθð Þ 1− cosθð Þ2 ð18Þ

The dependence of γeff/γ on the contact angle predicted usingEq. (18) is shown in Fig. 8. For poor wetting of the impurity by thenuclei (θ=180°), γeff=γ, and the impurity does not promote nu-cleation. For good wetting (θ=0°), the effective interfacial tensionis less than the melt-nuclei interfacial tension, and so the impurity

Fig. 8. Calculation of the normalized effective interfacial tension on the contact angle atthe solid–liquid interface. The heterogeneous nucleation rate increases as γeff/γ de-creases because then it is easier to form a nuclei in solution.


promotes nucleation leading to a faster nucleation rate. The value ofthe contact angle will depend on the molecular characteristics of theimpurity, solute, and solvent, e.g., their charge, polarities and hydro-phobicities. It may also be altered by the presence of any surface ac-tive substances in the various phases.

2.3.4. Nucleation promoters and inhibitorsIn the previous section, we discussed impurities present within a

melt or solution that could promote nucleation by forming surfacesthat reduced the free energy required for nuclei formation. In this sec-tion, we discuss other substances that are able to either promote orinhibit nucleation in melts or solutions. Incorporation of these sub-stances into a systemmay be useful for preventing undesirable nucle-ation and crystal growth of a bioactive component, or for controllingthe relative rates of nucleation and crystal growth [60]. A variety ofsubstances have been shown to be able to promote or inhibit nucle-ation, including cosolvents, surfactants and polymers [61]. A numberof physicochemical mechanisms have been proposed to account forthese effects [60]:

• Promotion: Surface-active additives (such as surfactants, phospho-lipids, or proteins) may reduce the effective interfacial tension(γSL) by adsorbing to the solid–liquid interface.

• Inhibition: Additives capable of forming micelles, vesicles or otherself-assembled colloidal structures (such as surfactants or phospho-lipids) may be able to incorporate hydrophobic solutes within thenon-polar regions of their internal structures. Consequently, themaximum concentration of solute that can be solubilized in theaqueous phase (CS⁎) would be increased, leading to a reduction inthe driving force for nucleation (CB/CS⁎).

• Inhibition: Additives with non-polar binding sites on their surfaces(such as globular proteins or cyclodextrins) may be able to bind hy-drophobic solutes, thereby increasing the maximum concentrationof solute that can be solubilized in the aqueous phase (CS⁎), againreducing the driving force for nucleation (CB/CS⁎).

• Inhibition: Additives that can increase the micro-viscosity of thecontinuous phase (such as sorbitol, glycerol or sugars) may slowdown the diffusion of solute molecules through the solution, there-by retarding their incorporation into the solid–liquid surface.

• Inhibition: Additives that are incorporated into or absorbed to thesolid–liquid interface may be capable of inhibiting the further incor-poration of solute molecules, thereby retarding nucleation.

Someof thesemechanisms change the thermodynamic driving force(ΔG⁎) for nucleation by altering the activation energy for nuclei forma-tion, whereas others change the kinetics of nucleation by altering therate of molecular diffusion or incorporation into the solid–liquid sur-face. A number of substances capable of inhibiting nucleation (andthereby promoting supersaturation) in pharmaceutical applicationshave been discussed in recent review articles, including certain surfac-tants, polymers, and cyclodextrins [60,62]. Some of these substancesor equivalent food-grade substances may be suitable for application infunctional food products to inhibit or control nucleation.

2.4. Crystal growth

Once stable nuclei have been formed in a melt or solution theygrow into crystals by incorporating additional solute molecules intothe solid–liquid interface [26,41]. Crystals have different faces, andeach face usually grows at a different rate, which partially accountsfor the wide variety of different crystal shapes that are formed. Theoverall crystal growth rate depends on a number of factors, includingmass transfer of the liquid molecules to the solid–liquid interface,mass transfer of non-crystallizing species away from the interface, in-corporation of the liquid molecules into the crystal lattice, or removalof the heat generated by the crystallization process from the interface.Any of these processes can be rate limiting depending on the molec-ular characteristics of the system and the prevailing environmentalconditions, e.g., temperature profile and mechanical agitation. Conse-quently, a general theoretical model of crystal growth is difficult toconstruct. In crystallizing lipid systems, the incorporation of a mole-cule at the crystal surface is often rate-limiting at high temperatures,whereas the diffusion of a molecule to the solid–liquid interface isoften rate-limiting at low temperatures [41]. This is because the vis-cosity of the liquid oil increases as the temperature is lowered andso the diffusion of a molecule is retarded. The crystal growth ratetherefore increases initially with supercooling, has a maximum rateat a certain temperature, and then decreases on further supercooling.The dependence of the growth rate on temperature therefore shows asimilar trend to the nucleation rate, however, the maximum rate ofnuclei formation usually occurs at a different temperature to the max-imum rate of crystal growth (Fig. 9). As a practical consequence, it ispossible to control the size of the crystals produced by manipulatingthe cooling conditions to control the relative rates of nucleation andgrowth [63]. If the system is supercooled to a holding temperaturewhere the nucleation rate is faster than the crystallization rate, thena large number of small crystals will be formed. On the other hand,if the system is supercooled to a holding temperature where the nu-cleation rate is slower than the crystallization rate, then a small num-ber of large crystals will be formed (Fig. 9). A variety of mathematicaltheories have been developed to model the rate of crystal growth incrystallizing materials [41,46], with the most appropriate model fora specific situation depending on the rate limiting step [41].

It should be noted that once crystallization is complete, there maystill be changes in crystal size and shape during storage due to post-crystallization processes, such as crystal aggregation or Ostwald rip-ening [26,41]. Crystal aggregation occurs when two or more crystalscome together and form a larger crystal, whereas Ostwald ripeningoccurs when oil molecules migrate from smaller crystals to large crys-tals through the intervening medium. Aggregation and Ostwald rip-ening therefore both lead to an increase in the average size of thecrystals present. Crystal growth during storage is often undesirablesince it adversely affects the physicochemical, functional and sensoryproperties of the final product.

0

20

40

60

80

100

0 10 20 30 40

Rel

ativ

e R

ate

Supercooling (ºC)

Nucleation

Crystallization

Rapid NucleationSlow Growth

Slow NucleationRapid Growth

TM

Fig. 9. Schematic representation of the dependence of the nucleation rate and crystalgrowth rate on temperature. Nucleation and crystal growth only occur once the mate-rial has been cooled below the melting point (TM).


Recently, a relatively simple mathematical model has been devel-oped to describe the growth of crystals once stable nuclei are formedin supersaturated solutions [36,53]. The rate of growth in the crystalradius with time is given by the following expression:

dRdt

¼ ψD0Vm

rCB−CS⁎ð Þ ð19Þ

Here, r is the crystal radius, t is the time, D0 is the diffusion coeffi-cient of the solute molecules through the liquid surrounding the crys-tals, Vm is the molar volume of the molecules forming the crystals, CBis the total solute concentration in the liquid surrounding the crystals(on a molar basis), and CS* is the solubility of the solute in this liquidat equilibrium (on a molar basis). The parameter ψ takes into accountthat crystal growth may be limited by different physicochemical phe-nomenon:

ψ ¼ rr þ λ

ð20Þ

Here λ is a “surface integration factor” which is a measure of theresistance of a solute molecule being incorporated into the crystalstructure. When λ≪r (ψ=1) the rate limiting step is diffusion of sol-ute molecules to the crystal surface, but when λ≫r (ψ=r/λ) the ratelimiting step is incorporation of solute molecules into the crystalstructure. The above equation shows that the crystal growth rateshould increase with an increase in solute diffusion coefficient (de-crease in solution viscosity) and with an increase in the total concen-tration of excess solute molecules present relative to the solubilitylimit. This equation describes how the size of a single crystal will in-crease over time after a nuclei has been formed. Information aboutthe number of nuclei formed per unit time per unit volume (the nu-cleation rate) is required to use this equation to simulate the crystalgrowth process [64].

2.5. Crystal morphology

The morphology of the crystals formed depends on a number ofinternal factors (e.g., molecular structure, composition, packing andinteractions) and external factors (e.g., temperature–time profile,

mechanical agitation, and impurities) [65]. As mentioned above,when a liquid lipid phase is cooled rapidly to a temperature wellbelow its melting point a large number of small crystals are formed,but when it is cooled slowly to a temperature just below its meltingpoint a smaller number of larger crystals are formed. This is becausethe nucleation rate increases more rapidly with decreasing tempera-ture than the crystallization rate (Fig. 9). Thus, rapid cooling producesmany nuclei simultaneously that subsequently grow into small crys-tals, whereas slow cooling produces a smaller number of nuclei thathave time to grow into larger crystals before further nuclei areformed. Crystal size has important implications for the stability, rhe-ology, sensory perception and biological fate of emulsion-based deliv-ery systems. When crystals are too large they are perceived as being“grainy” or “sandy” in the mouth [26]. Large crystals of a bioactivecomponent may dissolve more slowly and be absorbed more slowlyin the GIT. The efficiency of molecular packing in crystals also de-pends on the cooling rate. If a fat is cooled slowly, or the degree ofsupercooling is small, then the molecules have sufficient time to beefficiently incorporated into a crystal. At faster cooling rates, or higherdegrees of supercooling, the molecules do not have sufficient time topack efficiently before another molecule is incorporated. Thus, rapidcooling tends to produce crystals that contain more dislocations,and in which the molecules are less densely packed [66]. The coolingrate therefore has an important impact on the morphology and func-tional properties of crystals formed frommelts. Similarly, the size andmorphology of crystals formed in supersaturated solutions will de-pend on the degree of supersaturation of the system, as this will affectthe relative rates of nucleation and growth.

As mentioned earlier, crystals may come in a variety of differentmorphologies depending on the solute type, solvent type, and prepa-ration conditions, e.g., cuboids, spheroids, spherulites, and needles.Certain types of hydrophobic materials may also form fiber-like crys-talline networks within either the oil or aqueous phase, which canlead to the creation of a three-dimensional network that gives solid-like characteristics [67–69]. This type of behavior has been reportedfor phytosterols, which are crystalline hydrophobic bioactive mate-rials that have been reported to reduce cholesterol levels [70,71].

2.6. Polymorphism

Many hydrophobic bioactive substances exhibit a phenomenonknown as polymorphism, which is the ability of a material to exist in anumber of crystalline structures with different molecular conformationor packing [26,35,41,72,73]. Differences in the structural organization ofmolecules within different polymorphic forms leads to differences inthermodynamic (e.g., enthalpy, entropy, and free energy) and physico-chemical (e.g., melting point, heat of fusion, and density) properties.Usually one of the polymorphic forms has the lowest free energy andis therefore thermodynamically stable under a given set of conditions(Fig. 10). However, a lipidmay crystallize into a less-stable polymorphicform and remain there due to the fact that the activation energy for for-mation of this form is less than the most stable form (Fig. 11).

The number and types of polymorphic forms that a hydrophobicmolecule can exist in depends on its molecular properties and the en-vironmental conditions. For triacylglycerols, the three most common-ly occurring types are hexagonal, orthorhombic and triclinic, whichare usually designated as α, β′ and β polymorphic forms, respectively[73,74]. The thermodynamic stability of the three forms decreases inthe order: β>β′>α. Even though the β form is the most thermody-namically stable, triacylglycerols often crystallize in one of the meta-stable states because they have a lower activation energy of nucleiformation (Fig. 11). With time the crystals transform to the most sta-ble state at a rate that depends on environmental conditions, such astemperature, pressure, and the presence of impurities. The transfor-mation of a crystal from one polymorphic form to another can some-times be inhibited by adding substances that interfere with the

image of Fig.�9

G

Temperature

Amorphous

C

B

A

SupercooledLiquid

Liquid

Tm,C

Tm,A

Tm,B

Tt,B-C

Cooling

Tg

Fig. 10. Schematic representation of the temperature dependence of the free energy of alipophilic substance that can exist in different physical states: liquid, crystalline and amor-phous. In this example, the substance can crystallize into three different polymorphicforms (A, B and C) that have different melting temperatures (Tm), a solid–solid transition(Tt) can occur between the B and C polymorphic forms, and an amorphous form can beformed at the glass transition temperature (Tg). Adapted from Rodriguez-Sprong (2004).


process [75]. Polymorphic transitions often occur at a different rate inemulsified lipids than in bulk lipids (Section 3.2.2). The polymorphicform of a lipid influences the physicochemical properties and biolog-ical activity of many hydrophobic drugs (e.g., particle stability, solu-bility or dissolution rate) [35,75], and so it is often important tochoose preparation conditions that favor the formation and preserva-tion of a particular polymorphic state [76,77].

It should be highlighted that many hydrophobic materials canexist in an amorphous state as well as in a crystalline state dependingon their nature and the preparation conditions used. Amorphoussolids tend to have higher solubilities than their crystalline counter-parts, which may be an advantage for increasing the oral bioavailabil-ity of some hydrophobic components [35]. Consequently, it may beimportant to select preparation conditions that favor the formationof an amorphous state, rather than a crystalline state.

Liquid

Reaction Time

Energy Barriers:

Solid

G

GB

GA

GL

Molecular TransportMolecular AssemblyInterface Creation

Supercooling orSupersaturation

Driving Forces:

Fig. 11. Schematic representation of the kinetics of nucleation for two different poly-morphic forms (A and B) of a lipophilic substance. For the substance to convert fromthe liquid state to the solid state it must overcome a free energy (G) barrier, which isdifferent for different polymorphic forms. Thus a system can crystallize into a polymor-phic form that is not the most thermodynamically stable because the energy barrier isless. Adapted from Rodriguez-Sprong (2004).

3. Nucleation and growth in emulsions

The nature of the various physicochemical processes occurring inindividual lipid or aqueous phases (Section 2) may be altered appre-ciably when a lipid phase is emulsified with an aqueous phase to forman oil-in-water emulsion [25,78]. These effects can be associated withthe extremely small volume of the individual lipid phases (droplets),the presence of an interfacial layer around each droplet, and the par-titioning of bioactive components between the oil and aqueousphases. The influence of emulsification on the phase behavior of hy-drophobic bioactive components is briefly reviewed in this section.

3.1. Nucleation in emulsions

The formation of nuclei due to supercooling or supersaturation isoften appreciably different when it is emulsified than when it is inthe bulk form because of the extremely small volume of the lipidphase within each droplet [26,78]. For droplets with radii rangingfrom about 10 nm to 10 μm the droplet volume varies from about10−6 to 103 μm3 (Table 3) The formation of nuclei usually proceedsby a heterogeneous nucleation mechanism in a bulk liquid due to thepresence of catalytic impurities, such as dust, reverse micelles, andother substances [73]. On the other hand, the formation of nucleiwithin the lipid droplets in emulsified oil often takes place througha homogeneous nucleation process because the likelihood of finding acatalytic impurity within a particular droplet is very small (Fig. 12)[44,79]. Consequently, the lipid in an emulsion usually has to be super-cooled to a much greater degree than that in bulk oil before nucleationand crystal growth are observed. For example, the degree of supercool-ing required to promote nucleation has been reported to be typically 10to 20 °C less in emulsified oils compared to bulk oils [80]. The reason forthis phenomenon is the fact that any impurities present in the oil phaseare distributed between a huge number of individual droplets (Table 3),and nucleation in a particular droplet acts largely independently of nu-cleation in other droplets [44,49]. If a droplet contains an impurity, thennucleation will tend to occur due to heterogeneous nucleation, but if itcontains no impurity, then it will tend to occur through homogeneousnucleation [43]. This section primarily focuses on nucleation due tosupercooling, but many of the arguments also apply to nucleation dueto supersaturation.

3.1.1. Homogeneous nucleation within dropletsIf the total number of catalytic impurities present in an emulsion is

much lower than the total number of droplets present (NI≪ND), thenthe formation of nuclei will primarily be through homogeneous nucle-ation. Indeed, emulsions containing small droplets are commonlyused to study the kinetics of homogeneous nucleation in lipids [81].The rate of nuclei formation in an emulsion underdoing homogeneousnucleation can be described by the following equation [49]:

ΦN tð Þ ¼ 1− exp −Jtð Þ ð21Þ

Table 3Influence of particle size on particle properties. The calculations of the number of par-ticles per unit volume were made assuming that 1 cm3 of hydrophobic component washomogenized to form particles.

Droplet radius Droplet volume Number density Surface/volume(μm) (μm3) (m−3) (m−1)

0.01 4.2×10−6 2.4×1017 3×108

0.1 4.2×10−3 2.4×1014 3×107

1 4.2×100 2.4×1011 3×106

10 4.2×103 2.4×108 3×105

100 4.2×106 2.4×105 3×104

1000 4.2×109 2.4×102 3×103

Decreasing Particle Size: Increasing Droplet Number

NI >> ND NI << NDNI ND≈

Fig. 12. Schematic representation of the influence of particle size on the probability of finding an impurity within a given droplet. As the particle size decreases, the number of drop-lets (ND) increases relative to the number of impurities (NI), and therefore the fraction of droplets containing an impurity is reduced.


Where ΦN(t) is the fraction of droplets that contain a nuclei attime t, J is the nucleation rate (Section 2.3), and t is the incubationtime at a fixed holding temperature below the melting point. If it as-sumed that crystal growth occurs very rapidly within the dropletsdue to their small sizes, then the above equation can be used to pre-dict the change in the solid fat content with time: SFC(t)=1−ΦN(t)=exp(− Jt). The solid fat content of the lipid phase in an emul-sion should therefore increase exponentially with time until itreaches unity (when expressed as a mass fraction).

In practice, the increase in solid particles with time often does notfollow a single exponential function as predicted by this equation,with the rate of nucleation decreasing with increasing incubationtime [49]. This effect may be attributed to a number of causes:(i) the droplets in an emulsion are polydisperse, rather than mono-disperse [49]; (iii) the melting point of a material within a particle de-creases as the particle size decreases due to the increasing Laplacepressure, i.e., TM decreases with decreasing r [26]; (iii) there will bea distribution of catalytic impurities throughout the lipid phase sosome droplets may contain an impurity and some may not [49];(iv) impurities initially present within the lipid phase are expelledfrom the crystals formed, thereby increasing their concentration inthe non-crystalline phase, which leads to a decrease in the meltingpoint of the lipid [49]; (v) the lipid phase at the oil–water interfacehas a different molecular environment than that in the interior ofthe droplet, which may alter its melting point [43,82].

3.1.2. Volume heterogeneous nucleation within dropletsIf the number of catalytic impurities is much greater than the

number of droplets present in an emulsion (NI≫ND), then eachdroplet will contain an impurity and will tend to crystallize througha heterogeneous nucleation mechanism (Fig. 12). Typically, heteroge-neous nucleation occurs at an appreciably higher temperature thanhomogeneous nucleation. Indeed, it is often observed only a few de-grees below the melting point of the lipid phase. In this case, nucleistart to form on impurities distributed throughout the volume ofthe oil, and so it can be referred to as volume heterogeneous nucle-ation. As discussed previously the main factor influencing the rate ofnuclei formation due to the presence of catalytic impurities is theability of the lipid to wet the surface of the impurity, which dependson the contact angle (Section 2.3). If the wetting of the surface of animpurity by the nuclei is favorable, then the effective interfacial ten-sion will be reduced, thereby reducing the activation energy for nu-clei formation, and leading to an increase in nucleation rate.Consequently, nucleation is observed at higher temperatures in thepresence of catalytic impurities, i.e., less supercooling is required. Atpresent, the precise nature of the catalytic impurities present in dif-ferent kinds of lipids is usually unknown. They may be dust particlesfrom the atmosphere, association colloids in the lipid phase (such as

reverse micelles), or some other contaminants [73]. Volume hetero-geneous nucleation only tends to occur in emulsions when the num-ber of droplets per unit volume is relatively low or the number ofimpurities per unit volume is relatively high.

3.1.3. Mixed nucleation within dropletsWhen the number of impurities present within an emulsion is on

the same order of magnitude as the number of droplets present(NI≈ND), then a fraction of droplets will contain no catalytic impuri-ties while another fraction will contain one or more impurities(Fig. 12). In this situation, nucleation may occur through a combina-tion of homogeneous and heterogeneous nucleation. The fraction ofdroplets in a monodisperse emulsion containing impurities that arecapable of promoting heterogeneous nucleation can be described bythe following equation [26]:

ψI ¼ 1− exp −πd3NI

6

!ð22Þ

Where d is the droplet diameter and NI is the number of impuritiesper unit volume of oil capable of promoting nucleation. This equationassumes that the impurities are randomly distributed among thedroplets. Theoretical predictions of the influence of particle size andnumber of impurities per unit volume of oil on the fraction of dropletsthat contain an impurity are shown in Fig. 13. As the particle size in-creases, the fraction of droplets containing an impurity increases.Those particles containing an impurity would be expected to undergoheterogeneous nucleation, whereas those without an impurity wouldbe expected to undergo homogeneous nucleation.

An expression has been given for the fraction of droplets that arecrystalline at a particular temperature assuming that nucleationmay occur through either a heterogeneous or homogeneous nucle-ation mechanism [49]:

ΦN tð Þ ¼ 1−ΨI exp −JHettð Þ þ 1−ΨIð Þ exp −JHomtð Þ ð23Þ

Here JHet and JHom are the rates for heterogeneous nucleation andhomogeneous nucleation (Section 2.3), respectively. A number of ex-perimental studies have been performed using emulsions containingdifferent droplet sizes, and they have shown that it is possible to iden-tify two populations of droplets, one inwhich nucleation occurs rapid-ly (impurity-containing droplets) and the other where nucleationoccurs slowly (impurity-free droplets) [83]. Expressions have alsobeen derived to predict the fraction of droplets crystallizing as thetemperature is decreased at a controlled rate [49]. These equationsare useful for assessing the impact of cooling rates on the formationof crystals during preparation of emulsion-based delivery systems.

NI >> ND

NI << ND

Fig. 13. Dependence of the faction of droplets containing a catalytic impurity on parti-cle size and number of impurities present. Here ND is the number of droplets per unitvolume, and NI is the number of impurities per unit volume.


3.1.4. Surface heterogeneous nucleation within dropletsIn emulsified oils, it is possible for nucleation to occur at the oil–

water interface [84], rather than spontaneously within the droplet in-terior (homogeneous nucleation) or on the surfaces of impuritieswithin the droplets. Nucleation at the oil–water interface has been re-ferred to as surface heterogeneous nucleation to distinguish it fromthe nucleation that occurs at the surfaces of impurities distributedwithin the interior of the droplet which is usually referred to as vol-ume heterogeneous nucleation. A number of studies have shownthat the type of emulsifier present at the oil–water interface influ-ences the temperature at which nucleation is first observed in super-cooled oil-in-water emulsions [82,85–89]. Certain types of emulsifiersare able to promote nucleation at higher temperatures than others. Inparticular, surfactants that have a hydrophobic tail similar in molecu-lar structure to the nucleating oil have been proposed to increase thenucleation rate by acting as “templates” for nuclei formation [83,86].It may be possible to use this phenomenon to control the rate of nu-cleation and crystal formation by selecting an appropriate emulsifierto stabilize the lipid droplets.

3.1.5. Nucleation within the aqueous phaseSo far we have assumed that nucleation occurs with the lipid

droplets, but it is also possible for nucleation to occur within theaqueous phase [90]. This would occur if the hydrophobic substancehad a significant solubility in the aqueous phase, and the activationenergy for nuclei formation was lower in the aqueous phase than inthe oil phase. This activation energy is related to the interfacial ten-sion at the boundary separating the nuclei from the surrounding me-dium (Section 2.3.2), which may be different for nuclei formed in theoil and aqueous phases. For homogeneous nucleation, the hydropho-bic molecules within nuclei are surrounded by other lipid moleculesin the oil phase or by water molecules in the aqueous phase. Normal-ly, the interfacial tension will be higher for the nuclei-water boundarythan for the nuclei-oil boundary due to the hydrophobic effect, andtherefore this phenomenon would tend to favor formation of nucleiin the oil phase (lower activation energy). However, the aqueousphase usually contains surface-active materials (emulsifiers) thatcould adsorb to the surfaces of the nuclei or crystals formed, whichmay alter the interfacial tension. In addition, there will also be differ-ences in the concentrations of the hydrophobic molecules in the oiland aqueous phases (N0), as well as the viscosities of the oil and aque-ous phases, which both influence the rate that hydrophobic

molecules come together and form clusters. There may also be differ-ent types and levels of impurities in the oil and aqueous phases,which would alter the relative rates of heterogeneous nucleation inthe two phases. At present, little is understood about the origin of nu-cleation in oil-in-water emulsions, and of the consequences for thestability and functional performance of emulsion-based delivery sys-tems. Nevertheless, some progress has been made in mathematicalmodeling of nucleation and crystal growth in emulsions [90]. Thismodel shows that nucleation can begin in either phase dependingon the nature of the system.

3.2. Crystal growth and properties within emulsions

Once nuclei have been formed in a particular location within anemulsion they may grow into crystals. The size, shape, and locationof crystals in an emulsion will affect its physical stability and func-tional performance [25]. Hydrophobic molecules move through thesurrounding liquid until they encounter a growing crystal surface,where they may be incorporated into the crystal structure. Crystalscontinue to grow until all of the hydrophobic molecules have been in-corporated or the saturation limit is reached (CB=CS⁎), after whichthe formation of new crystals will be stopped. Nevertheless, theremay still be changes in the size, shape and polymorphic form of crys-tals during storage due to molecular rearrangements and mass trans-port phenomenon [41,49].

3.2.1. Crystal size and morphologyThe size and shape of the crystals present within an emulsion-

based delivery systemwill influence its physical stability (e.g., particleaggregation and gravitational separation) [91,92], its physicochemicalproperties (e.g., rheology, appearance, and mouthfeel) [93], and its bi-ological activity (e.g., aggregation state, dissolution and absorptionrates in the GIT) [37,94–96]. It is therefore important to understandthe physicochemical factors that influence the size and morphologyof the crystals formed within emulsions, so that these factors can becontrolled to produce desirable crystal characteristics for specificapplications.

In general crystal properties will depend on the nature of the hy-drophobic component, the composition of the oil and aqueous phases,the initial emulsion preparation conditions, and the subsequent stor-age conditions. Intuitively, one might expect that the size of the crys-tals present in an emulsion containing a hydrophobic bioactivecomponent would be limited by the size of the droplets. If the hydro-phobic component crystallizes within the droplets, then there is aphysical boundary to crystal growth due to the presence of the oil–water interface. In principle, crystals with different sizes can then beproduced by preparing oil-in-water emulsions with different initialdroplet sizes [96]. In practice, the relationship between crystal anddroplet characteristics depends on the nature of the crystallizing sub-stance [90]. For certain lipid types it is possible to produce crystallineparticles of different sizes using emulsion droplets as templates[8,97]. For example, solid triacylglycerol or hydrocarbon particleswith different sizes can be produced by cooling liquid dropletsbelow their melting points [45,98,99]. Nevertheless, there is oftenan appreciable change in the shape of the particles formed upon crys-tallization of the dispersed phase. For example, when solid triacylgly-cerol particles are formed they may initially be spherical, but theymay quickly change to ellipsoid or other shapes [93,100,101]. Lipiddroplets have recently been used as templates for producingmicron-sized cholesterol crystals [102]. For other lipid types, the rela-tionship between the size of the crystals present in an emulsion andthe initial size of the oil droplets is more complex. A recent studyfound that β-carotene crystals dispersed in aqueous solutions growduring storage because of Ostwald ripening, i.e. diffusion of solutemolecules from small to large crystals [103]. Recent experiments inour laboratory have shown that crystals that are much larger than

Fig. 14. Overall appearance and polarized microscopy images of corn oil-in-wateremulsions stabilized by Tween 20 in the presence of polymethoxyflavones (PMF) mea-sured after 1 day storage. The size of the crystals formed (L>10 μm) are much largerthan the initial droplet size (db1 μm). Images taken from Li et al. [37].


the initial droplet size can be produced for certain types of hydropho-bic components, e.g., flavonoids (Fig. 14). At present we still have afairly poor understanding of the physicochemical phenomenon oc-curring during nucleation and crystal growth in these multiphasesystems.

3.2.2. Mass transport processesThe size and morphology of the crystals formed within an

emulsion-based delivery system may change over time due to masstransport processes, such as Ostwald ripening (OR) [104]. In bulklipids, OR occurs due to molecular diffusion of hydrophobic moleculesthrough the oil phase and leads to a net increase in the overall size ofthe crystals present. In emulsified lipids, the OR process is more com-plicated since crystals may be present in both the oil and aqueousphases. Consequently the hydrophobic molecules may diffusethrough both the oil and aqueous phases leading to transfer of a ma-terial from one phase to another. In addition, mass transport process-es may also be driven by differences in the interfacial tensions at thesolid–liquid boundaries of the crystals in the oil and aqueous phases.There will be a tendency for hydrophobic molecules to move fromcrystals with high interfacial tensions to those with lower interfacialtensions.

There is currently a relatively poor understanding of changes incrystal morphology due to mass transport processes in multi-phasesystems such as oil-in-water emulsions. Nevertheless, some insightsinto the factors that influence this process can be obtained by exam-ining the theory of Ostwald ripening for spherical particles. The in-crease in mean particle radius with time due to Ostwald ripening isgiven by [26]:

r3−r30 ¼ 8γSLVm

9RTCS⁎Dt ð24Þ

Here Vm is the molar volume of the solute, γSL is the interfacialtension at the crystal–liquid interface, CS⁎ is the equilibrium solubilityof the solute in the surrounding liquid, r0 is the initial particle radius,D is the translation diffusion coefficient of the solute through the con-tinuous phase, R is the gas constant, and T is the absolute tempera-ture. These equations indicate that the cube of the mean particlesize should increase linearly with time, interfacial tension, and solutesolubility. One would therefore expect crystal growth in emulsionsdue to mass transfer processes to increase as the solubility of the hy-drophobic material in the oil and aqueous phases increased, as the

viscosities of the two phases decreased, or as the interfacial tensionsincreased.

3.2.3. PolymorphismAs mentioned earlier, many hydrophobic materials can exist in

more than one crystalline (or amorphous) state when they are cooledbelow their melting point (Section 2.6). The polymorphic state that isformed immediately after cooling is usually the one that has the low-est activation energy, which may not be the form with the lowest freeenergy (Fig. 11). Consequently, the crystals may undergo a transitionfrom the initial metastable polymorphic form to a more stable poly-morphic form during storage. For example, triacylglycerols willoften initially crystallize into the α-polymorphic form, but then trans-form into the β-polymorphic form during storage [98,101,105]. Therate of polymorphic transformations is usually much more rapid inemulsified lipids than in bulk lipids. For example, studies with triacyl-glycerols have shown the transformation from the α-to-β polymor-phic occurs much faster in the emulsified form than in the bulkform [106]. This effect has been attributed to the smaller crystalsizes in emulsified fats compared to bulk fats [106], but it may alsobe due to interfacial phenomenon.

A change in the polymorphic state of an emulsified hydrophobiccomponent may have a major impact on the physicochemical proper-ties of emulsion-based delivery systems [98,101,105]. For example,when liquid triacylglycerol droplets are initially cooled well belowtheir melting point they crystallize into the α-polymorphic form,which leads to the formation of spheroid solid particles with diame-ters similar to that of the initial droplets. However, when these solidparticles undergo a α-to-β polymorphic transformation the particleschange into an ellipsoid shape, which leads to extensive particle ag-gregation and system gelation (Fig. 15). Particle aggregation hasbeen attributed to the large increase in particle surface area that oc-curs when the particles are transformed from a spherical to a non-spherical shape [93,107,108]. The surface area of the particles isthen not completely covered by emulsifier molecules, which leadsto aggregation through hydrophobic patches on particles surfaces. Anumber of approaches have been developed to either retard polymor-phic transformations or to prevent subsequent particle aggregation:(i) control of storage temperature [105]; (ii) addition of oil-solubleinhibitors (such as some surfactants) that retard polymorphic trans-formations [76,109]; (iii) addition of excess emulsifier to cover hy-drophobic patches formed on particle surfaces [108]; (iv) utilizationof emulsifiers that increase the repulsive interactions between parti-cles [108].

3.2.4. Crystal locationThe crystals in an emulsion may be present in the oil phase, the

aqueous phase, or at the oil–water interface depending on their phys-icochemical properties [25]. The location of the crystals may have animportant impact on the physical stability and functional perfor-mance of an emulsion-based delivery system. The location of crystalsin an emulsion depends on their “wettability”, which is a measure oftheir relative affinity for the oil and water phases [25]. Wettability isusually characterized in terms of the contact angle (θ), which is gov-erned by the interfacial tensions at the various boundaries in a threecomponent system, in this case solid–water, solid–oil and water–oil.

cosθ ¼ γOS−γWS

γOWð25Þ

The magnitude of these interfacial tensions depends on the rela-tive strength of the molecular interactions between the molecules atthe boundaries, e.g., van der Waals, hydrogen bonding, hydrophobicinteractions, and electrostatic interactions [110]. The interfacial ten-sions and contact angle therefore depend on the molecular character-istics of the phases that make up the system (e.g., their polarity,

Solid ( -Form) Solid

( -Form)Liquid

Spheroid SpheroidPlatelet

Stable Stable

Aggregation

αβ

Fig. 15. Lipid particles often undergo appreciable changes in their morphology and stability after crystallization and polymorphic transformations.


charge, and hydrophobicity). In practice, there may also be sub-stances present in the oil or water phases that adsorb to the variousboundaries present in the system and alter the interfacial tensions,such as emulsifiers (e.g., surfactants, phospholipids, proteins, or poly-mers). The presence of these surface-active agents may thereforealter the physical location of the crystals in an emulsion, whichcould alter its physical stability (through partial coalescence).

The contact angle (θ) is usually specified through the waterphase, i.e., a tangent is drawn to the surface of the solid particleand the angle between this line and the oil–water interface is mea-sured through the water phase (Fig. 16). Thewetting of the solid par-ticle by the water phase increases as the contact angle decreases(Fig. 17), which influences the location of the crystals (Table 4),such as the distance they protrude into the aqueous phase(Fig. 16). So far we have assumed that the crystals are spherical par-ticles with smooth surfaces, but in reality they are usually non-spherical particles with irregular surfaces, which makes the analysisof their location more complex [25].

If crystals form within the interior of a droplet, then they maymove to the droplet surface by diffusing through the oil phase. The

Fig. 16. The location of a particle at a boundary between an oil and aqueous phase canbe described by the contact angle (θ) and the distance that it products into the waterphase (δ).

average time for crystals to move from the droplet interior to theoil–water interface can be described by the following equation [25]:

tI ¼6πη0rcr

2d

kBTð26aÞ

Here, η0 is the viscosity of the oil phase, rC is the radius of the crys-tal, rD is the radius of the droplet, and kBT is the thermal energy. Anindication of the time required for crystals to reach the oil–water in-terface can be obtained by using representative values: η=50 mPas(rC=100 nm, rD=1000 nm): tI=25 s. In practice, crystals may beunable to move to the droplet surface, even though they may be bet-ter wetted by water, because a crystal network forms within thedroplet that prevents their movement. An approximate estimationof the time required for network formation is given by the followingequation [25]:

tN ¼ 3η04kBTn0

ð26bÞ

Where n0 is the number of crystals per unit volume. If it is as-sumed that the crystals are spherical, then this equation can be writ-ten as:

tN ¼ πη0r3c

ϕkBTð27Þ

Where ϕ is the volume fraction of crystals present within the lipidphase. An estimate of whether the crystals can reach the interface canbe given by the ratio tI/tN:

tItN

¼ 6ϕr2cr2d

ð28Þ

If tI/tN≪1, then the crystals should reach the interface beforeforming a network within the droplets, but if tI/tN≫1, the crystalswill form a network before reaching the interface and so they willtend to remain within the droplet interior. This equation suggeststhat the smaller the size of the crystals relative to the droplets, themore likely that a network will form within the droplets. It also high-lights that the lower the crystal concentration (ϕ) within the oil

Fig. 17. The wetting of a solid particle (crystal) by the water phase increases as the contact angle (θ) decreases, which depends on the balance of interfacial energies in the system(solid–oil, solid–water, oil–water).


phase, the more chance they have at reaching the droplet surfacewithout aggregating. In practice, there will also be a minimum crystalconcentration required for network formation. If the crystal concen-tration is below this value, the crystals may aggregate but they maystill be able to move to the droplet surface. This critical concentration(c⁎) will depend on the shape of the crystals and the strength of theattractive forces between them. For fat crystal networks, typically 5to 10% crystals are required to form a three-dimensional networkwith solid-like properties [26].

3.3. Influence of crystals on emulsion stability

The presence of hydrophobic crystals in oil-in-water emulsionsmay have a pronounced influence on their physicochemical proper-ties and stability [25,91,109,111]. When the droplets are partiallycrystalline, a crystal from one droplet can penetrate into anotherdroplet during a collision which causes the two droplets to stick to-gether. This phenomenon is known as partial coalescence and leadsto particle aggregation. Particle aggregation may also occur whencrystalline lipid particles undergo morphological changes(Section 3.2.2), which has been attributed to the exposure of hydro-phobic patches at the particle surfaces after the liquid–solid phasetransition [108]. In dilute systems aggregation leads to an increasein viscosity and a decrease in creaming stability, whereas in concen-trated systems aggregation may lead to gelation [105,107,112]. Onemust therefore be careful when designing emulsion-based deliverysystems that any crystals present (bioactive components or carrierlipid) do not promote physical instability. A number of approacheshave been developed to improve the aggregation stability of emul-sions containing crystals [20,25,26]: (i) ensure the fat crystalsremain within the droplets and do not protrude into the aqueousphase; (ii) ensure there is enough emulsifier present to completelycover all of the particle surfaces; (iii) coat the particle surfaces withan interfacial layer that prevents the droplets from coming closeenough to aggregate; (iv) ensure that the emulsions are not subjectedto mechanical agitation, since this can promote droplet aggregation[111].

Table 4Relationship between contact angle of a solid particle (crystal) at an oil–water inter-face, and its location in an oil-in-water emulsion.

Contact angle Preferential wetting Crystal location

0° Water Water phase45° Water Interface, protruding into water phase90° Water and Oil Interface135° Oil Interface, protruding into oil phase180° Oil Oil phase

4. Encapsulation of hydrophobic bioactive agents in emulsions

There have been appreciable technical and conceptual advancesin our ability to produce emulsions over the past few years, whichhave been reviewed recently [6,21,113,114]. In general, methodsused to fabricate emulsion-based delivery systems can be catego-rized into two broad categories depending on their underlying prin-ciple: high-energy methods and low-energy methods [22,114–116]High-energy approaches utilize mechanical devices (“homoge-nizers”) capable of generating intense forces that disrupt and inter-mingle the oil and water phases leading to the formation of smalloil droplets dispersed in water, e.g., high shear mixers, colloid mills,high pressure valve homogenizers, microfluidizers, and sonicationmethods [1,19,115,117,118]. High-energy methods are widely usedin the food, pharmaceutical and other industries to prepare emul-sions because they are versatile, relatively robust, and capable oflarge-scale production. Low energy approaches rely on the sponta-neous formation of small oil droplets within mixed oil–water-surfac-tant systems when either the composition and/or environmentalconditions are altered, e.g., spontaneous emulsification, phase inver-sion, and emulsion inversion point methods [22,113,119–122]. Atpresent, low energy methods are mainly used in the pharmaceuticalindustry, but some of them may also be suitable for applications inthe food industry. The minimum particle size that can be producedusing each approach depends on a number of factors includingthe system composition and the operating conditions for the specifichomogenization method used [123]. A number of applications ofemulsion-based delivery systems for hydrophobic bioactive compo-nents are shown in Table 5.

4.1. Incorporation into the oil phase

The first step in encapsulating a hydrophobic component intoan emulsion is to incorporate it into the oil phase prior to homog-enization. Liquid bioactive components (such as ω−3 oils or Vita-min E-Acetate) may be used as the oil phase themselves, or theymay be incorporated into a carrier lipid by simple mixing. On theother hand, solid bioactive components (such as carotenoids, cur-cuminoids or phytosterols) usually have to be either melted or dis-solved prior to homogenization since crystalline materials cannotbe emulsified directly using most high or low energy methods. Inthese cases, it is necessary to either dissolve or melt the crystallinematerial prior to homogenization, which was discussed inSection 2.1 for bulk oils and is reviewed briefly here for oil phasesthat will be homogenized. If the carrier lipid matrix phase is to besolidified, as in solid lipid nanoparticles (SLN) or nanostructuredlipid carriers (NLC), then these materials should also be meltedprior to homogenization.

Table 5Examples of recent studies using emulsion-based delivery systems for encapsulating crystalline hydrophobic bioactive components.

Bioactive component Carrier oil Emulsifier Factors References

Lycopene Olive oil Tomato pulp Storage temperature [185]Sunflower oil+MCT Tween 20 (neutral) Chelating agents, Antioxidants [186]Sunflower oil+hexadecane SDS (−) Light, pH, chelating agents, antioxidants, transition metals [148]Sunflower oil+hexadecane,corn oil or stripped corn oil

SDS (−), DTAB (+), polyoxyethylene laurylether (neutral)

Oil type, surfactant type, fatty acid addition [187]

Astaxanthin Organic solvents Sodium Caseinate Homogenizer pressure, number of passes [188]MCT Tween 20, WPI Emulsifier type, Oil content, homogenization conditions [189]

β-Carotene MCT Soybean soluble polysaccharides and chitosan Interfacial properties, storage temperature [190]MCT Modified starch, whey protein, Tween 20,

monoglyceridesHomogenization conditions, Emulsifier Type [191]

MCT or tripalmitin Lecithin, Tween 20, Tween 80 Carrier oil type, emulsifier type [149]Organic solvent (acetone) Sodium caseinate, sucrose ester, Tween 20,

monoglyceridesEmulsifier Type [192]

Vegetable oil Sugar ester, gelatin Emulsifier type, homogenization method [193]Organic solvent (hexane) Sodium caseinate, Whey protein Emulsifier Type [194]MCT Tween 20, 40, 60 and 80 Emulsifier type and concentration, homogenization

conditions, storage temperature[195]

Organic solvent (hexane) Tween 20 Homogenizer shear rate and time [196]Lutein Marigold oil Gum arabic, sucrose ester Bioavailability [197]

MCT Labrasol Oil, water, surfactant composition [198]Corn oil Whey protein, phospholipids System composition, homogenization conditions [199]Triglyceride oil Phospholipids Carotenoid structure, In vitro digestion and transfer [200]

Curcumin MCT Tween 20 Particle size [201]MCT, trimyristin, tristerin,glyceryl monosterate

Tween 80, Poloxamer Emulsifier type, Carrier oil type [202]

CoEnzyme-Q – Phospholipid Delivery system type [203]Triglyceride oil Lecithin Delivery system type [204]Olive, safflower, coconut,butter and cocoa oils

Lecithin, monoglycerides, stearoyl-2-lactate,DATEM

Emulsifier Type, Carrier oil type [124]

Resveratrol Fish oil and tributyrin Milk protein, modified starch In vivo bioavailability [205]Coconut oil Lecithin, non-ionic surfactants Delivery system formulation [206]Coconut oil, Vitamin E Lecithin, non-ionic surfactants Delivery system formulation [207]

Phytosterols Organic solvent (hexane) Tween 20 Homogenization pressure [208]Corn oil Brij 35 Phytosterol structure (cholesterol, stigmasterol,

β-sitosterol, 5α-cholestane)[209]


4.1.1. MeltingIn principle, a solid lipid can be converted into a liquid lipid by

heating it above its melting point (Section 2.1.1). An emulsion canthen be formed by homogenizing this hot liquid lipid phase witha hot aqueous emulsifier solution, making sure that the tempera-ture remains above the melting point of the lipid throughout thehomogenization procedure. In reality, this approach is only suitablefor bioactive components that are chemically stable at elevatedtemperatures, and whose melting point is not too high for commer-cial homogenizers to operate at. After the emulsion is preparedat high temperatures it can be cooled below the crystallizationpoint of the bioactive component to promote the formation ofsolid particles. The resulting colloidal dispersion will consist ofemulsifier-coated crystalline lipid particles dispersed within anaqueous medium. The properties and stability of these kinds of sys-tem will depend on the nature of the crystals formed, e.g., polymor-phic form, size, and shape. If the melting point of the bioactivecomponent is above about 100 °C then it is necessary to use highpressures to avoid water evaporation during homogenization.

4.1.2. DissolutionFrom a thermodynamic point of view, a crystalline material has

a finite solubility in an oil phase (CS⁎)—below this level the materi-al fully dissolves, but above this level it forms crystals. In practice,it is usually possible to dissolve a greater amount of a crystallinematerial into an oil phase than CS⁎ due to supersaturation. A solu-tion becomes supersaturated because of the activation energy as-sociated with the formation of nuclei required to grow crystals(Section 2.1.2). To completely avoid the possibility of crystalliza-tion during the manufacture, storage, or utilization of a deliverysystem it is necessary to ensure that the solute concentration al-ways remains below the saturation level. This limits the total

amount of solute that can be incorporated into a delivery system.If it is assumed that the hydrophobic bioactive component (thesolute) is only soluble in the oil phase, then the maximum amountof solute that can be incorporated into the system to completelyavoid crystallization is:

Φ⁎SO ¼ C⁎

SOΦO

1−C⁎SO

ð29Þ

WhereΦSO⁎ is themass fraction of the solute in the overall system atsaturation,ΦO is the mass fraction of the oil phase (carrier oil+solute)present in the system, and C⁎SO is the saturation concentration of thesolute in the oil phase (expressed as a mass fraction). This equationcan be used to calculate whether an emulsion-based delivery systemcan actually be formulated that contains a sufficiently high amountof bioactive component to have a biological effect. For example, if itis known that a human should consume 10 mg per day of a bioactivecomponent to obtain a biological effect, then one can calculatewhether this amount can be incorporated into a commercial productusing an emulsion-based system. Assuming the emulsion-based de-livery system contained 10 wt.% oil phase (ΦO=0.1) and C⁎SO was0.1 wt.% (C⁎SO=0.001) for the bioactive in question (i.e., 1 mg of bio-active component per 1 g of oil phase), then one would have to con-sume approximately 100 g of the delivery system per day to reachthe required level. Some bioactive components may partition intoboth the oil and water phases, and therefore a more sophisticatedanalysis is required (Section 5).

The saturation concentration of a bioactive component dependson the nature of the solvent (oil phase) used. Recent studies in ourlaboratory showed that the saturation concentration of curcumin de-creased in the following order for long, medium and short chain


triglycerides: SCT>MCT>LCT. Studies with CoEnzyme Q (CoQ10)

found that its solubility depended on the nature of food-grade triglyc-erides oils used, e.g., MCT, olive, safflower, coconut, butter and cocoaoils [124,125]. It may therefore be possible to increase the amountof a bioactive component that can be encapsulated in an emulsion-based delivery system by careful selection of carrier oil.

4.1.3. Melting–dissolutionIt is also possible to use a combination of dissolution and melt-

ing to produce emulsion-based delivery systems containing encap-sulated hydrophobic components. In this case the crystallinematerial is dispersed in carrier oil, which is then heated to dissolvethe crystals. An oil-in-water emulsion is then formed by homoge-nization of the liquid oil and water phases together in the presenceof a water-soluble emulsifier. The advantage of this method is thatconsiderably lower temperatures can be used to melt all of thecrystalline material. Nevertheless, crystals may form in the emul-sion during storage when the system is cooled [37]. This approachwas discussed in Section 2.1.3 and the change in solubility of arepresentative bioactive component with temperature is shownin Fig. 4.

4.2. Encapsulation using high energy methods

High energy methods use mechanical devices to generate intenseforces that intermingle and disrupt oil–water-emulsifier mixtures,leading to the formation of small droplets of one of the phases dis-persed within the other phase. A number of different homogenizersare commonly used in industry to produce emulsions (Fig. 18), in-cluding high shear mixers, high pressure homogenizers, colloidmills, ultrasonic homogenizers, and microfluidizers [20,26,126]. The

Fig. 18. Various mechanical devices that can be used to produce emulsions using a high-eultrasonic jet homogenizer; ultrasonic probe homogenizer; colloid mill; membrane homo

choice of a particular type of homogenizer and of the specific operat-ing conditions used depend on the characteristics of the materialsbeing homogenized (e.g., viscosity, interfacial tension, shear sensitiv-ity) and the required final emulsion properties (e.g., droplet size,droplet concentration, viscosity). A crystalline bioactive lipid wouldnormally be dispersed in the oil phase prior to homogenization by en-suring that it is below its saturation concentration in the carrier oil orby warming the lipid phase to melt any crystals present. If the bioac-tive lipid was susceptible to chemical degradation (e.g., carotenoids),then it may be necessary to carefully control homogenization condi-tions to avoid exposure to factors that increase the degradation rate(e.g., high temperatures, oxygen, light or transition metals) or toadd chemical protectants (such as antioxidants, chelating agents orbuffers).

In general, the particle size produced by high energy approaches isgoverned by a balance between two opposing processes occurringwithin the homogenizer: droplet disruption and droplet coalescence[127]. The smallest droplet size that can be produced by a particularhigh-energy device depends on homogenizer design (e.g., flow andforce profiles), homogenizer operating conditions (e.g., energy inten-sity, duration), environmental conditions (e.g. temperature, pressure),sample composition (e.g., oil type, emulsifier type, concentrations),and the physicochemical properties of the component phases (e.g.,interfacial tension, viscosity) [117,128]. Previous studies have shownthat the droplet size tends to decrease when the energy intensity or du-ration increases, the interfacial tension decreases, the emulsifier adsorp-tion rate increases, and the disperse-to-continuous phase viscosity ratiofalls within a certain range (0.05bηD/ηCb5) [22,26,126]. The extent ofthe ηD/ηC range where small droplets can be produced depends on thenature of the disruptive forces generated by the particular homogenizerused, i.e., simple shear versus extensional.

nergy approach: high shear mixer; high pressure valve homogenizer; microfluidizer:genizer. Adapted from McClements (2011).


High energy approaches are one of the most versatile means ofproducing emulsions industrially because they can be used with awide variety of different oil and emulsifier types, and are suitablefor large scale production. Provided that the homogenization condi-tions are suitably optimized, emulsions can be produced using triacyl-glycerol oils, flavor oils, and essentials oils as the oil phase, as well aswith proteins, polysaccharides, phospholipids, and surfactants as theemulsifier. Even so, the size of the particles produced depends strong-ly on the characteristics of the oil and emulsifier used [4,6]. For exam-ple, it is usually easier to produce very small droplets when the oilphase has a low viscosity and/or interfacial tension (e.g., flavor oils,essential oils, alkanes) than when it has a high viscosity and/or inter-facial tension (e.g., MCT or LCT). The principles behind high energymethods used to produce emulsions have been reviewed elsewhere,e.g., high shear mixers, colloid mills, high pressure homogenizers,microfluidizers, ultrasonic homogenizers, and membrane methods[6,20,123].

4.3. Encapsulation using low energy methods

The formation of emulsions using low-energy approaches relieson the spontaneous formation of oil droplets in oil–water-surfactantmixtures when either their composition or their environment is al-tered [22,113,114,119–122]. A number of commonly used approachesfor producing emulsions using low energy methods are shown inFig. 19. Low energy approaches are sometimes more effective at pro-ducing small droplet sizes than high energy approaches, but they areoften more limited in the types of oils and emulsifiers that can beused. For example, it is currently difficult to use polymer-based emul-sifiers (such as proteins or polysaccharides) in most of the low energyapproaches used to form emulsions. Instead, it is often necessary touse relatively high concentrations of synthetic small molecule surfac-tants to form emulsions by these approaches, which may limit their

Oil phase:Oil + Water-

soluble Surfactant

Aqueous phase:Water

O/W Emulsion

W/O Emulsion

Bicontinuous

Cool

Phase Inversion Temperature

Spontaneous Emu

Fig. 19. Examples of low-energy methods that can be used to produce oil-in-water emulsioversion temperature; emulsion inversion point. Adapted from McClements (2011).

use for some commercial applications. There has been considerableprogress recently in rationalizing the formation of emulsions by lowenergy methods using a conceptual approach that describes thephase behavior of surfactant-oil–water (SOW) mixtures [129,130].The principles behind low energy methods used to produce emul-sions have been reviewed elsewhere, e.g., spontaneous emulsification,phase inversion temperature, phase inversion composition and emul-sion inversion point methods [4,6,21,22,113,129,131–133].

5. The partitioning of bioactive substances in emulsions

Once an oil-in-water emulsion containing a hydrophobic bioactivecomponent has been fabricated it is possible for the physical state andlocation of the bioactive component to change [25,90]. Many crystal-line hydrophobic bioactive components have a finite solubility in bothoil and aqueous phases. Consequently, when they are dispersed intoan emulsion-based delivery system they may be distributed betweenthe oil and aqueous phases (if they are below their saturation concen-tration), or some of them may even be present as crystals (if they areabove their saturation concentration). In this section, we consider thepartitioning of crystalline hydrophobic components in oil-in-wateremulsion systems, and highlight some of the major factors that influ-ence partitioning.

At relatively low concentrations, the solute (hydrophobic compo-nent) will be fully dissolved in both the oil and water phases, and itsdistribution between the phases will be determined by the equilibri-um partition coefficient (KOW). At relatively high concentrations, thesolute will be above its saturation concentration in the oil (CSO⁎)and water (CWO⁎) phases and so it will tend to form a separate solidphase (Fig. 20). Practically, one would like to know the maximumamount of solute that can be fully dissolved in the oil–water systemat equilibrium, as well as the concentrations of solute in the oily, wa-tery, and solid phases.

lsification

Emulsion Inversion Point

ns based on of spontaneous droplet formation: spontaneous emulsification; phase in-

image of Fig.�19

Oil PhaseMass = MO + MSO

Solubility = CSO*

Water PhaseMass = MW + MSW

Solubility = CSW*

Solid PhaseMass = MS

KOW = CSO / CSW

Fig. 20. A solute will partition between the oily, watery and solid phases depending onthe oil–water partition coefficient and the equilibrium solubilities. Key: the symbols Mand C refer to mass and solubility concentration, whereas the subscripts S, O and Wrefer to solute, oil and water, respectively.

Fig. 21. Influence of equilibrium partition coefficient (KOW) on the maximum amountof solute that can be solubilized within an emulsion system at saturation as the oil con-tent is increased. For lipophilic components (KOW>1), the amount that can be incorpo-rated within an emulsion increases with oil content. It is assumed that the saturationconcentration of the solute in oil (CSO*) is 0.05.

Fig. 22. Influence of oil content and equilibrium partition coefficient (KOW) on the per-centage of a solute that is solubilized within the oil phase on an emulsion-based deliv-ery system. For more non-polar compounds, a greater percentage is present within theoil phase than for more hydrophilic compounds.


5.1. Maximum amount of solute that can be dissolved in oil–watersystem

A mass-balance analysis of the distribution of the solute in an oil–water system leads to the following expression for the maximumamount of solute that can be dissolved at equilibrium saturation:

Φ⁎S ¼

C⁎SO ΦO KOW−1ð Þ þ 1−C⁎

SO

� �h ��

1−C⁎SO

KOW

ð30Þ

Here, CSO⁎ (=MSO⁎/[MSO⁎+MO]) and CSW⁎ (=MSW⁎/[MSW⁎+MW])are the equilibrium saturation concentrations of the solute in the oiland water phases (expressed as mass fractions),ΦO is the mass frac-tion of oil in the overall system (=MO/ [MS+MW+MO]), and KOW isthe oil–water partition coefficient. MSO and MSW are the masses ofthe solute dissolved in the oil and water phases, while MS, MO andMW refer to the masses of the solute, oil and water, respectively.The superscript “⁎” refers to the condition of saturation. This equa-tion can be used to calculate the total amount of solute dissolved inan oil–water system as a function of system composition (ΦO) andsolute properties (CSO⁎ and KOW). Predictions of the influence of par-tition coefficient and oil content on this value are shown in Fig. 21.For hydrophobic solutes (KOW>1), the amount of solute that canbe incorporated within an emulsion increases with oil content.

5.2. Solute partitioning below the saturation level

At solute concentrations below the saturation level, the solutemolecules are fully dissolved in both phases with a distribution deter-mined by the partition coefficient. The fraction of the solute mole-cules present in the oil phase is given by the following equation:

ϕSO ¼ −b−ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffib2−4ac

p2a

!ð31Þ

Where: a=[1−KOW]; b=[KOW(ΦS−ΦO)+(ΦO−1)]; c=KOW

ΦSΦO. A plot of the percentage of the solute dissolved in the oilphase (100×ΦSO/ΦS) as a function of oil phase concentration isshown in Fig. 22 for substances with different oil–water partition co-efficients. As expected, the more hydrophobic a material (higherKOW), the higher the fraction that remains within the oil dropletswhen the oil concentration is changed. For less hydrophobic mate-rials (lower KOW), a significant fraction of the substance may be

present within the aqueous phase, particularly when the oil concen-tration is relatively low. This may have important consequences forthe functional performance of many emulsion-based delivery sys-tems. First, the chemical stability of many hydrophobic componentsdepends on their molecular environment, e.g., some active sub-stances undergo faster degradation when dissolved in water thanwhen dissolved in oil [134]. Second, the absorption of hydrophobicsubstances from the gastrointestinal tract will depend on whetherthey are dissolved in the aqueous phase or present within lipid drop-lets. Third, nucleation, crystal growth and ripening effects may de-pend on the molecular environment of the hydrophobic substance.

image of Fig.�22


5.3. Solute partitioning above the saturation level

Above the saturation level for the system (ΦS>ΦSO⁎), the solutewill be saturated in both the oil phase (CSO⁎) and the water phase(CSW⁎), and a third solid phase will be formed (usually crystals)(Fig. 20). The concentration (mass fraction) of the solute in the oil,water and solid phases is then given by:

Φ⁎SO ¼ C⁎

SOΦO

1−C⁎SO

Φ⁎SW ¼ C⁎

SOΦW

KOW−C⁎SO

Φ⁎SS ¼ ΦS−Φ⁎

SO−Φ⁎SW ð32Þ

Here, ΦS is the total mass fraction of solute present in the overallsystem. These equations can be used to determine the change in con-centration with system composition (ΦS, ΦO, ΦW) and properties(CSO⁎ and KOW). Calculations of the solute concentration in the threedifferent phases with increasing solute concentration in the overallsystem are shown in Fig. 23. Initially, when the solute is added itonly partitions between the oil and water phases, but once a criticallevel is exceeded it forms a separate solid phase. Under these circum-stances there is a possibility of crystals forming within the emulsionduring storage, which could alter the physical stability of the deliverysystem, as well as the bioavailability of the solute.

5.4. Influence of carrier lipid properties on solute partitioning

The nature of the carrier oil usedwithin an emulsion-based deliverysystemwill influence solute partitioning and crystallization. The chem-ical composition and physicochemical properties of the carrier oil influ-ence the maximum amount of solute that can be solubilized in the oilphase (CSO⁎) as well as the oil–water partition coefficient (KOW).

So far we have considered the partitioning of solute molecules inan emulsion that contains liquid oil and aqueous phases. Neverthe-less, it is also possible for the carrier lipid to be fully or partially crys-talline, which can have a major impact on solute partitioning. It hasbeen proposed that highly hydrophobic substances can be trappedwithin colloidal crystalline matrices (solid lipid nanoparticles) so asto improve their chemical stability during storage and to controltheir release kinetics [97,135,136]. It is postulated that the hydropho-bic material is trapped within the crystalline matrix, which slows

Saturation

Fig. 23. Calculated concentration of solute in oil, water and solid phases when the massfraction of solute in the system is increased. It was assumed that: CSO*=0.01, KOW=2,ϕW=0.8.

down its diffusion as well as the diffusion of any reactants thatmight promote instability (such as oxygen). In practice, it is oftenfound that once a carrier lipid crystallizes the solute molecules are ex-pelled because they cannot easily be incorporated into the highly or-dered structure of the crystalline phase. As a result the net amount ofthe liquid oil phase in the system is reduced, which will alter the rel-ative concentration of the solute in the oil and water phases. The vol-ume of the oil phase will be reduced to a value of (1−SFC) of itsoriginal value, where SFC is the solid fat content or the mass fractionof the total carrier lipid phase that is solid. The equations given inSections 5.1 to 5.3 would therefore have to be modified to take thisphenomenon into account. A number of experimental studies haveobserved this effect in practice, e.g., partitioning of aroma compounds[137,138], flavor compounds [139], and unsaturated oils [140], anddrugs [97,136].

5.5. Influence of particle size on partitioning

The partitioning of a bioactive component in an emulsion systemmay also depend on the size of the particles. For example, the equilib-rium solubility of the material in a spherical particle increases as thesize of the particles decreases [60]:

C rð Þ ¼ CS⁎ exp2γSLVm

RTr

� �ð33Þ

Here Vm is the molar volume of the solute, γSL is the interfacialtension at the solid–liquid interface, CS⁎ is the equilibrium solubilityof the solute in the continuous phase, and C(r) is the solubility ofthe solute when contained in a spherical particle of radius r. Thus ifthe bioactive component makes up the entirety of the lipid phase inan emulsion its partitioning may depend on particle size. A plot ofthe change in relative solubility (C(r)/CS*) of a solute on particlesize is shown in Fig. 24 for a representative bioactive component(β-carotene). For relatively large particles (r>1000 nm), the solubil-ity of the solute in the surrounding liquid is close to the solubility of abulk material. However, as the particle radius decreases below about200 nm there is a steep increase in the solubility of the material in thesurrounding liquid. The effect of particle size on solubility has beendemonstrated experimentally for various lipophilic bioactive compo-nents [60,178].

Fig. 24. Dependence of the solubility of a solute on particle size calculated usingEq. (33). The solubility increases appreciably when the particle radius falls below about200 nm. The calculations were made for β-carotene particles, with molar volume of570 cm3 mol−1, and an assumed interfacial tension of 1 mJ m−2.

image of Fig.�23


6. Chemical stability of encapsulated bioactive components

A number of hydrophobic bioactive components are chemically un-stable and will tend to degrade over time, e.g., polyunsaturated lipids,carotenoids, curcuminoids, flavonoids, and phytosterols [141–143].The mechanism and rate of chemical degradation of these componentswithin an emulsion may be quite different from that of the same com-ponent dispersed within a pure oil phase, because of the presence ofan oil–water interface, an aqueous phase, and the various oil-soluble,water-soluble and amphiphilic functional components used in emul-sions [142]. For unstable bioactive components it is important to estab-lish the main factors responsible for their chemical degradation so thatthese factors can be controlled during the preparation and storage of theemulsion-based delivery system. These factors vary from system to sys-tem depending on the molecular structure of the bioactive componentand its environment. In this section, the potential role of lipid crystalli-zation on the chemical stability of hydrophobic bioactive componentsin emulsion-based delivery systems is highlighted using carotenoidsas an example.

Carotenoids are a group of isoprene-based hydrophobic compo-nents that typically have 40-carbon molecules and multiple conjugateddouble bonds [144]. The potential for carotenoids for promoting healththrough a number of physiological mechanisms has been shown, in-cluding preventing oxidative damage, quenching singlet oxygen, alter-ing transcriptional activity, and serving as precursors for vitamin A[141,144–147]. Because they contain multiple conjugated doublebonds carotenoids are highly susceptible to chemical degradation,which leads to color fading and loss of biological activity. For example,the rate of lycopene degradation has been shown to increase with de-creasing pH, increasing temperature, exposure to light, and the pres-ence of singlet oxygen [148]. A number of studies have examined theincorporation of carotenoids into emulsions, and its impact on theirchemical degradation (Table 1) [141]. The crystallization of the emulsi-fier layer (lecithin) or the carrier oil (tripalmitin) has been found to in-hibit the chemical degradation of β-carotene in oil-in-water emulsions,which was attributed to their ability to trap the carotenoids in a solidmatrix that prevented them from coming into close proximity to pro-oxidants in the aqueous phase [149]. On the other hand, a recentstudy found that crystallization of the carrier oil (hydrogenated palm

Tight junctions

Para-cellular

Large particles may be unable to enter the

mucous layer.

Nanoparticles may be digested differently

Fig. 25. Small particles may be trapped within the mucous layer, or may be transportedsubstances.

kernel oil) in oil-in-water emulsions actually promoted the chemicaldegradation of encapsulated β-carotene [150]. This effect can be attrib-uted to the ability of the carrier oil to form a highly regular crystallinestructure that expels the carotenoid from the interior of the droplets,thereby exposing it to pro-oxidants in the aqueous phase. These studieshighlight the importance of correctly designing an emulsion-based de-livery system so that the carrier lipid forms a solid matrix that traps,rather than expels, the encapsulated labile bioactive component.

7. Biological fate and bioavailability of encapsulated crystallinematerials

The successful design of emulsion-based delivery systems for en-capsulation and release of hydrophobic bioactive components de-pends on understanding their fate within the gastrointestinal tract(GIT) once they have been ingested [5,143,151–156]. This section fo-cuses on some of the ways that emulsion-based delivery systems mayalter the biological fate of hydrophobic bioactive compounds, withspecial emphasis on the role of lipid crystallization. The lipid phasethat forms crystals may be the bioactive component itself or it maybe the surrounding lipid matrix (carrier oil).

7.1. Behavior of emulsion-based delivery systems within the GI tract

7.1.1. Ingestion and digestionA brief overview of the complex conditions existing in different lo-

cations of the GI tract is given in this section (Figs. 25 and 26), withspecial emphasis on their influence on the properties of emulsion-based delivery systems [153,157,158].

7.1.1.1. Mouth. After ingestion an emulsion-based delivery system ex-periences various physiological and physiochemical conditions with-in the oral cavity: it is mixed with saliva and undergoes changes inpH, ionic strength and temperature; it is exposed to digestive en-zymes; it interacts with the surfaces of the tongue, mouth and throat;it experiences complex flow/force profiles during ingestion, mastica-tion and swallowing [159]. As a result the lipid droplets may aggre-gate with each other, coat the surfaces of the oral cavity, or becomemixed with mucin from the saliva. Any crystals present within an

Mucous layer

Nanoparticles may be trapped in the mucous layer, which increases their retention time.

Trans-cellular

Lumen

across it can be adsorbed directly, which could alter the bioavailability of lipophilic

image of Fig.�25

Enzyme Adsorption& Activity

00.20.40.60.8

1

0 100 200 300Fra

ctio

n R

elea

sed

Time (s)

Matrix Disruption

Droplet Formation & Disruption

Competitive Adsorption& Displacement

Solubilization & Transport

Accumulation at site of action

Mouth• pH 5-7• Enzymes• Salts• Biopolymers• 5 – 60 s

Stomach• pH 1-3• Enzymes• Salts• Biopolymers• Agitation• 30 min – 4 hours

Small Intestine• pH 6-7.5• Enzymes• Salts, Bile• Biopolymers• Agitation• 1 – 2 hours

Colon• pH 5-7• Enzymes• Bacteria• Agitation• 12-24 hours

Crystal precipitation & dissolution

Fig. 26. Highly schematic diagram of the various physicochemical and physiological processes that may occur when emulsion-based delivery systems pass through the human GItract. Picture of human body was obtained from http://en.wikipedia.org/wiki/Digestive_tract (Copyright free).


ingested emulsion may be altered due to their presence within theoral cavity, and in turn may influence the behavior and fate of the de-livery system within the mouth. Crystals with a sufficiently low melt-ing point will melt in the mouth, which can impart a desirable coolingsensation due to the heat absorbed by the endothermic solid–liquidphase transition [26]. The melting of lipid crystals may also alter thepartitioning of bioactive components in emulsion-based delivery sys-tems, e.g., if a bioactive component is embedded within a crystallinematrix, then it may be released when the matrix melts. Crystalswith a sufficiently high water-solubility may dissolve in the mouthwhen they are diluted by saliva, thereby altering the location of thebioactive components in the system. Relatively large individual crys-tals or crystal aggregates (>50 μm) may be detected as particles inthe mouth, which may impart an undesirable mouthfeel, e.g., a grittyor sandy perception.

7.1.1.2. Stomach. The material leaving the mouth (“bolus”) passesthrough the esophagus and enters the stomach where it may residefor periods ranging from a few minutes to a few hours dependingon its dimensions and physicochemical properties. Within the stom-ach the ingested material is exposed to highly acidic conditions(pH 1 to 3), is mixed with enzymes (e.g., proteases and gastric li-pases) and surface active substances (e.g., phospholipids and pro-teins), and, experiences a complex flow/force profile [160,161].Endogenous (from biological fluids) and exogenous (from food andits digestion products) surface active substances may displace theoriginal surface active materials from the lipid droplet surfacesthrough competitive adsorption processes. Gastric lipases initiatesome limited digestion of the lipid phase, while gastric proteases ini-tiate digestion of any absorbed and non-absorbed proteins. The con-ditions that the delivery system experiences within the stomachmay alter the nature of any crystalline material present. Crystals

may dissolve due to dilution in gastric juices or they may movefrom one location to another due to changes in contact angle broughtabout by alterations in the interfacial tensions at the crystal–oil, crys-tal–water, and oil–water interfaces. If a bioactive component has ion-izable groups, then its electrical characteristics may change when it ispresent in the acidic environment of the stomach, which may alter itssolubility in the surrounding aqueous medium (ionized solutes areusually more water-soluble than non-ionized ones).

Any crystals remaining in the stomach may alter the microstruc-ture and physicochemical properties of the gastric contents [162].Crystals may promote droplet aggregation through flocculation andpartial coalescence, which can alter the ability of gastric enzymes toaccess the digestible material within the aggregates [163]. Particlestrapped in the interior of flocs may be less accessible to digestive en-zymes than non-flocculated particles so that the subsequent lipid di-gestion rate is decreased. Droplet coalescence leads to a decrease inthe overall surface area of the lipid phase exposed to the digestive en-zymes (lipase), which may alter the digestion rate. Droplet aggrega-tion may also alter the spatial distribution of the material within thestomach through creaming or sedimentation, which could alter thegastric emptying rate and thereby the feeling of satiety [164–166].

7.1.1.3. Small intestine. The material entering the small intestine fromthe stomach (“chyme”) is mixed with alkaline digestive juices (con-taining bile salts, phospholipids, pancreatic lipase, colipase and bicar-bonate), which increases the pH close to neutral. Surface activesubstances present in the small intestine compete with the surfaceactive materials already present at the lipid droplet surfaces, andmay displace them thereby altering the droplet interfacial composi-tion. Various digestive enzymes may hydrolyze the components pre-sent within a lipid droplet: pancreatic lipase/co-lipase complexconvert triglycerides and diglycerides into monoglycerides and free

http://en.wikipedia.org/wiki/Digestive_tract


fatty acids; phospholipases convert phospholipids to free fatty acidsand other hydrophobic substances; proteases convert proteins topeptides and amino acids; amylases convert complex carbohydratesto sugars. Lipid digestion products and encapsulated hydrophobicbioactive components may be incorporated into mixed micelles(bile salt/phospholipid micelles and vesicles), and then transportedthrough the mucous layer to the surfaces of the enterocytes wherethey are absorbed (Figs. 25 and 26). The rate and extent of incorpora-tion of hydrophobic substances into mixed micelles depends on theirmolecular characteristics and physical state, as well as the nature ofthe surrounding food matrix.

Any crystals present in the chyme usually have to be dissolved inthe intestinal fluids before they can be incorporated into mixed mi-celles and absorbed by epithelium cells [27,167]. The dissolutionrate depends on many factors, including the solubility of the crystal-line material in the intestinal fluids and the crystal dimensions. Theextent of absorption also depends on the solubilization capacity ofthe mixed micelles, which depends on the type and amount of lipidsconsumed. Typically, the solubilizing capacity of mixed micelles in-creases as the amount of triacylglycerols consumed increases, sincethen there are more fatty acids and monoacylglycerols that canform part of the mixed micelles [16,17]. In addition, the solubilizationcapacity of mixed micelles tends to increase as the molecular weightof the fatty acids in the system increases, e.g., LCT>MCT>SCT [16,17].The physical state of a carrier lipid may also influence its rate of diges-tion, e.g., recent studies have shown that solid triglyceride particlesare digested at a slower rate than liquid triglyceride particles [168].This phenomenon may be useful for controlling the release of encap-sulated bioactive components within the GI tract. It may also be pos-sible for very fine crystalline particles to be directly absorbed byepithelium cells through a variety of transport mechanisms (seeSection 7.1.2).

The presence of crystalline material in the small intestine may alsoalter the microstructure and physicochemical properties of the othercomponents in the chyme [162]. Crystalsmay promote droplet aggrega-tion through flocculation and partial coalescence, which can alter theability of intestinal enzymes to access the digestible material trappedin the aggregates (e.g., lipids, proteins, and carbohydrates). A recentstudy using Raman scattering micro-spectroscopy showed that thepresence of a crystalline bioactive component (ergosterol) within lipiddroplets altered the digestion rate of the triglyceride carrier oil [167].The same study showed that the lipid digestion products formed bythe action of lipase on triglycerides were able to solubilize the bioactivecomponent into mixed micelles and other colloidal structures, whichshould increase its absorption and bioavailability. As lipid digestion pre-ceded the concentration of hydrophobic bioactive components in theundigested lipid phase increased, suggesting that they accumulated inthat phase prior to incorporation into mixed micelles. A similar findingwas recently reported for the release of β-carotene from oil-in-wateremulsions under simulated GI conditions [169].

As a result of exposure to these various physiological conditions,the droplets in emulsion-based delivery systems may change appre-ciably as they pass through the GI tract [170,171]. For example,there may be changes in the composition, size distribution, aggrega-tion state, electrical charge, interfacial properties, and physical stateof the droplets (Fig. 26). In addition, there may be appreciablechanges in the concentration, physical state, polymorphic form, andmorphology of any crystalline bioactive material present [27]. All ofthese changes need to be taken into account when trying to establishthe potential biological fate of emulsion-based delivery systems con-taining crystalline bioactive components.

7.1.1.4. Colon. Bioactive components that are not absorbed in thesmall intestine will eventually reach the colon, which may be eitherdesirable or undesirable. If the bioactive component has to beabsorbed into the systemic circulation before it can exhibit its

beneficial attributes, then any non-absorbed bioactive componentsreaching the colon would reduce the bioavailability. On the otherhand, if the bioactive component has a positive effect on the colon(e.g., it is a pre-biotic or anti-colon cancer component), then it maybe beneficial to have it reach the large intestine.

The colon typically has a pH close to neutral, but becomes mildlyacidic when dietary fibers are broken down to short chain fattyacids. Food material may remain in the large intestine for manyhours (typically 6–24 h), whichmay lead to further changes in the na-ture of any crystalline materials present, e.g., dissolution, aggregation,or absorption. Alternatively, the crystallinematerial may pass throughthe entire large intestine and eventually be excreted in the feces withthe other waste matter. In this case, a portion of the ingested bioactivematerial is not utilized by the body, which would reduce the bioavail-ability and bioactivity of the material. The large intestine has a largepopulation of endogenous bacteria, which may utilize any crystallinematerial as an energy source during fermentation, thereby leading tochemical alterations in the bioactive component. The resultingmetab-olites may have different physicochemical properties and biologicalactivities to the original ingested crystalline material.

7.1.2. AbsorptionOnce the lipid droplets and/or their digestion products reach the

surfaces of the GI tract they may be absorbed by various kinds of ep-ithelium cells [116,156,172]. Small lipid droplets or crystals may beabsorbed through a number of different mechanisms [174–177]:

− Paracellular mechanism: Sufficiently small particles may pass be-tween the narrow gaps (“tight junctions”) separating neighboringepithelial cells (Fig. 25). This mechanism has been suggested for thetransport of solid lipid nanoparticles across epithelium cells [176].

− Transcellular mechanism: Small crystals may be absorbed directlythrough epithelial cell membranes by either passive or activetransport mechanisms (Fig. 25). Typically, absorption occurs byan “endocytosis” mechanism that involves the particle encounter-ing the cell membrane, the membrane wrapping itself around theparticle, and then part of the membrane budding-off to form avesicle with a particle trapped inside.

The pathway taken, and the rate and extent of crystal uptake, de-pends on the properties of the particles involved, e.g., size, shape,charge and interfacial chemistry [156,173]. The physicochemicalcharacteristics of particles also determines their fate once they haveentered the epithelium cell: (i) they may be digested by cellular en-zymes into their constituent parts which may then be absorbed; (ii)they may be transported through the cell and out into the blood orlymph systems; or (iii) they may accumulate within specific locationsin the cell [156,174]. Particles that are transported out of the epithe-lial cells via the portal vein or lymphatic system may circulatethrough the human body, where they may then be metabolized, ex-creted, or accumulate within certain tissues [154]. .

One would expect that the direct absorption and accumulation oflipid droplets in emulsion-based delivery systems would be unlikelybecause they would be expected to be rapidly digested within thestomach, small intestine, or epithelium cells. However, if the lipiddroplets were formed from indigestible oils (such as hydrocarbonsor mineral oils) or if the droplets were coated with indigestible shells(such as dietary fibers) then direct absorption could occur. On theother hand, direct absorption of undissolved crystalline bioactivecomponents might occur if they were small enough to be absorbedby the paracellular or transcellular mechanisms mentioned above.

7.2. Factors affecting biological fate of bioactive components

7.2.1. Transit timeThe time that a lipid droplet spends within the GI tract may affect

the rate and extent of the enzymatic hydrolysis of digestible carrier


lipids (such as triacylglycerols), and therefore the dissolution, solubi-lization, and absorption of any encapsulated hydrophobic substances.Particle characteristics may influence the transit time of encapsulatedcomponents with the GIT through a number of mechanisms [154]:(i) small droplets may be able to penetrate into the mucous layercoating the enterocytes, whereas large droplets cannot (Fig. 25);(ii) cationic droplets may bind to the mucous layer (which consistsof anionic biopolymers) through electrostatic interactions. An in-crease in transit time should enable more of the carrier lipids to bedigested, and a greater fraction of the bioactive component to be sol-ubilized by micelles and absorbed by enterocytes [116].

7.2.2. Carrier oil digestionLipid digestion occurs at the oil–water interface surrounding the

emulsified lipid droplets and so the rate of lipid digestion tends to in-crease as the droplet size decreases, because this increases the surfacearea of lipid exposed to the digestive enzymes [163,178]. Thereforeone would expect an emulsion-based delivery system containingsmaller droplets to generate lipid digestion products that can solubi-lize bioactive components more rapidly, thereby leading to a fasterrelease of encapsulated bioactive components. Nevertheless, therate of lipid digestion may not always decrease for smaller dropletsizes. Recent experiments in our laboratory found that globularprotein-stabilized nanoemulsions prepared using a homogenization/solvent evaporation method were digested at a slower rate than con-ventional emulsions [179], which was attributed to the fact that thenanoemulsions had a thicker protein coating than the conventionalemulsions. In addition, the fraction of non-adsorbed surfactanttends to increase as the droplet size increases, which can retardlipid digestion since then there is more free surfactant present tocompete at the oil–water interface with lipase [180].

The interfacial layer coating the oil droplets may also influence therate and extent of lipid digestion. A coating that prevents lipase fromadsorbing to the oil droplet surfaces will retard lipid digestion[180–182]. The composition of the carrier lipid may also affect therate and extent of digestion and absorption. Triacylglycerol oils(such as corn oil or fish oil) should be fully digested within the stom-ach and small intestine, whereas fat substitutes (such as Olestra)should not be. Thus, more of a bioactive may be released from triacyl-glycerol oils than from fat substitutes. Lipid digestion also depends onthe molecular characteristics of digestible triacylglycerols oils, e.g.,medium chain triglycerides (MCT) tend to be digested more rapidlythan long chain triglycerides (corn oil) [17,183]. Finally, liquid carrieroils tend to be digested more rapidly than solid carrier oils [168],which may alter the rate of mixed micelle formation and solubiliza-tion of bioactive components.

7.2.3. PrecipitationStudies in the pharmaceutical industry have shown that some hy-

drophobic drugs encapsulated within emulsion-based delivery sys-tems may precipitate when they are subsequently exposed tosimulated gastrointestinal conditions [27]. The same study showedthat the drug precipitated into an amorphous (rather than crystal-line) form during digestion, which would alter is solubility in mixedmicelles. The tendency for precipitation to occur and the nature ofthe solid phase formed has been shown to depend on the nature ofthe carrier oil and surfactant used to formulate emulsion-baseddrug delivery systems [184]. The possibility of precipitation occurringwithin the GI tract after consumption of crystalline hydrophobic bio-active food components (nutraceuticals) clearly requires furtherstudy, as it may have an important impact on their bioavailability.

7.2.4. DissolutionAbsorption of hydrophobic bioactive components by the human

body usually requires that they exist in a dissolved form (ratherthan solid form) within the small intestine. The rate and extent of

dissolution of crystalline hydrophobic components within the GItract is therefore an important factor influencing their bioavailability.The dissolution rate will depend on the nature of the solid particlesthemselves (crystalline versus amorphous, polymorphic form, crystalmorphology, and surface area), as well as the nature of the surround-ing intestinal fluids (volume, pH, composition, fluid flow). The disso-lution rate will tend to be faster for: amorphous than crystallinestates; for less stable polymorphic forms; for crystals with higher sur-face areas; for larger fluid volumes; for more intense mixing condi-tions. The influence of pH and ionic composition will depend on thetype of crystals present and of other charged components thatmight influence their dissolution rate (e.g., minerals, surfactants orpolymers). An increasing concentration of mixed micelles will tendto increase the dissolution rate by removing some of the free bioac-tive component from the aqueous phase (see next section).

7.2.5. SolubilizationThe particle characteristics within emulsion-based delivery sys-

tems may also influence the solubilization of hydrophobic bioactivecomponents within mixed micelles in the small intestine [17]. First,the presence of the lipid carrier itself stimulates the release of bilesalts and phospholipids from the bile duct, which increases the num-ber of endogenous mixed micelles present to solubilize the hydro-phobic component. Second, the surface-active digestion productsfrom the lipid carrier material, such as FFAs, MAGs and DAGs, are in-corporated into mixed micelles thereby increasing their solubilizationcapacity. A recent study of the structural changes occurring duringlipid digestion reported that a variety of other colloidal structuresmay also be formed that can solubilize bioactive components, includ-ing microemulsions and liquid crystalline structures [210]. Conse-quently, the bioavailability of hydrophobic bioactive componentscan be increased appreciably by controlling the amount and type ofcarrier lipid used to solubilize them.

7.2.6. AbsorptionThe relative bioavailability of highly hydrophobic compounds has

been shown to increase appreciably when the particle size is de-creased below about 500 nm [116]. The precise physicochemicalmechanism underlying this process will depend on the nature of thebioactive component involved. Some of the mechanisms that havebeen proposed to account for the increased absorption of bioactivecomponents in small particles:

(i) The water-solubility of highly hydrophobic compounds in-creases with decreasing particle size;

(ii) Small particles can diffuse into the mucous layer coating theenterocytes, which increases their retention time and theirability to interact with the enterocytes;

(iii) Very small particles may be able to cross the epithelia by di-rect absorption into enterocytes through passive or activetranscellular routes or by passing between cells by a paracel-lular route;

(iv) When the carrier oil is digested by lipase a dissolved hydro-phobic component may exceed its solubility limit and formcrystals. The size of the crystals formed will decrease as thelipid droplet size decreases, which may have important conse-quences for the absorption of the hydrophobic component.

There may also be components present within the delivery systemthat can enhance the absorption of a highly hydrophobic substance bypromoting active or passive transport mechanisms. For example, cer-tain kinds of polymers and surfactants are believed to open the tight-junctions between epithelium cells and thereby increase absorptionof very fine particles.


8. Conclusions

This review article has highlighted a number of the most impor-tant physicochemical and physiological processes associated withthe encapsulation and delivery of crystalline hydrophobic compo-nents using emulsion-based delivery systems. In particular, ithighlighted the importance of establishing the maximum amount ofhydrophobic component that can be successfully incorporated into adelivery system without crystallization occurring. It also highlightedsome of the major factors influencing the formation of crystals inemulsions, and their potential impact on the absorption of hydropho-bic components within the GI tract. This information is important forthe rational design of emulsion-based delivery systems for nutraceu-tical and pharmaceutical components.

Acknowledgments

This material is based upon work supported by the CooperativeState Research, Extension, Education Service, United States Depart-ment of Agriculture, Massachusetts Agricultural Experiment Stationand a United States Department of Agriculture, CREES, NRI and AFRIgrants.

References

[1] Velikov KP, Pelan E. Colloidal delivery systems for micronutrients and nutraceu-ticals. Soft Matter 2008;4:1964–80.

[2] Evans DF, Wennerstrom H. The colloidal domain: where physics, chemistry, biol-ogy, and technology meet Second ed. New York: Wiley-VCH; 1999.

[3] Madene A, Jacquot M, Scher J, Desobry S. Flavour encapsulation and controlledrelease—a review. Int J Food Sci Technol 2006;41:1–21.

[4] McClements DJ. Edible nanoemulsions: fabrication, properties, and functionalperformance. Soft Matter 2011;7:2297–316.

[5] McClements DJ, Decker EA, Weiss J. Emulsion-based delivery systems for lipophi-lioc bioactive components. J Food Sci 2007;72:R109–24.

[6] McClements DJ, Rao J. Food-grade nanoemulsions: formulation, fabrication,properties, performance, biological fate, and potential toxicity. Crit Rev FoodSci Nutr 2011;51:285–330.

[7] Flanagan J, Singh H. Microemulsions: a potential delivery system for bioactives infood. Crit Rev Food Sci Nutr 2006;46:221–37.

[8] Muller RH, Gohla S, Keck CM. State of the art of nanocrystals—special features,production, nanotoxicology aspects and intracellular delivery. Eur J Pharm Bio-pharm 2011;78:1–9.

[9] Chen L, Remondetto GE, Subirade M. Food protein-based materials as nutraceu-tical delivery systems. Trends Food Sci Technol 2006;17:272–83.

[10] Kesisoglou F, Panmai S, Wu YH. Application of nanoparticles in oral delivery ofimmediate release formulations. Curr Nanosci 2007;3:183–90.

[11] Kosaraju SL. Review of polysaccharides for encapsulation and delivery. Crit RevFood Sci Nutr 2005;45:251–8.

[12] Sagalowicz L, Leser M, Watzke HJ, Michel M. Monoglyceride self-assembly struc-tures as delivery vehicles. Trends Food Sci Technol 2006;17:204–14.

[13] Flangan J, Singh H. Microemulsions: a potential delivery system for bioactives infood. Crit Rev Food Sci Nutr 2006;46:221–37.

[14] Spernath A, Aserin A. Microemulsions as carriers for drugs and nutraceuticals.Adv Colloid Interface Sci 2006;128:47–64.

[15] Taylor TM, Davidson PM, Bruce BD, Weiss J. Liposomal nanocapsules in food sci-ence and agriculture. Crit Rev Food Sci Nutr 2005;45:587–605.

[16] Pouton CW. Formulation of poorly water-soluble drugs for oral administration:physicochemical and physiological issues and the lipid formulation classificationsystem. Eur J Pharm Sci 2006;29:278–87.

[17] Porter CJH, Trevaskis NL, Charman WN. Lipids and lipid-based formulations: op-timizing the oral delivery of lipophilic drugs. Nat Rev Drug Discov 2007;6:231–48.

[18] Porter CJH, Wasan KM. Lipid-based systems for the enhanced delivery of poorlywater soluble drugs. Adv Drug Deliv Rev 2008;60:615–6.

[19] McClements DJ. Emulsion design to improve the delivery of functional lipophiliccomponents. Annu Rev Food Sci Technol 2010;1:241–69.

[20] McClements DJ. Food emulsions: principles, practice, and techniques. 2nd ed.Boca Raton: CRC Press; 2005.

[21] Mason TG, Wilking JN, Meleson K, Chang CB, Graves SM. Nanoemulsions: forma-tion, structure, and physical properties. J Phys Condens Matter 2006;18:R635–66.

[22] Tadros T, Izquierdo R, Esquena J, Solans C. Formation and stability of nano-emulsions. Adv Colloid Interface Sci 2004;108–09:303–18.

[23] Quemada D, Berli C. Energy of interaction in colloids and its implications in rhe-ological modeling. Adv Colloid Interface Sci 2002;98:51–85.

[24] Pal R. Rheology of simple and multiple emulsions. Curr Opin Colloid Interface Sci2011;16:41–60.

[25] Rousseau D. Fat crystals and emulsion stability—a review. Food Res Int 2000;33:3–14.

[26] Walstra P. Physical chemistry of foods. New York, NY: Marcel Decker; 2003.[27] Sassene PJ, Knopp MM, Hesselkilde JZ, Koradia V, Larsen A, Rades T, et al. Precip-

itation of a poorly soluble model drug during in vitro lipolysis: characterizationand dissolution of the precipitate. J Pharm Sci 2010;99:4982–91.

[28] Donsi F, Wang YW, Li J, Huang QR. Preparation of curcumin sub-micrometer dis-persions by high-pressure homogenization. J Agric Food Chem 2010;58:2848–53.

[29] Vekilov PG. Nucleation. Cryst Growth Des 2010;10:5007–19.[30] Delaney JS. Predicting aqueous solubility from structure. Drug Discov Today

2005;10:289–95.[31] Modarresi H, Conte E, Abildskov J, Gani R, Crafts P. Model-based calculation of

solid solubility for solvent selection—a review. Ind Eng Chem Res 2008;47:5234–42.

[32] Wang JM, Hou TJ. Recent advances on aqueous solubility prediction. Comb ChemHigh Throughput Screen 2011;14:328–38.

[33] Ran YQ, He Y, Yang G, Johnson JLH, Yalkowsky SH. Estimation of aqueous solubil-ity of organic compounds by using the general solubility equation. Chemosphere2002;48:487–509.

[34] Yang G, Ran YQ, Yalkowsky SH. Prediction of the aqueous solubility: comparisonof the general solubility equation and the method using an amended solvationenergy relationship. J Pharm Sci 2002;91:517–33.

[35] Rodriguez-Spong B, Price CP, Jayasankar A, Matzger AJ, Rodriguez-Hornedo N.General principles of pharmaceutical solid polymorphism: a supramolecularperspective. Adv Drug Deliv Rev 2004;56:241–74.

[36] Lindfors L, Forssen S, Westergren J, Olsson U. Nucleation and crystal growth in su-persaturated solutions of a model drug. J Colloid Interface Sci 2008;325:404–13.

[37] Li Y, Zheng JK, Xiao H, McClements DJ. Nanoemulsion-based delivery systems forpoorly water-soluble bioactive compounds: influence of formulation parameterson polymethoxyflavone crystallization. Food Hydrocolloids 2012;27:517–28.

[38] Li X, Jastic BR. Design of controlled release drug delivery systems. New York, NY:McGraw-Hill Professional; 2005.

[39] Chebil L, Humeau C, Anthoni J, Dehez F, Engasser JM, Ghoul M. Solubility of fla-vonoids in organic solvents. J Chem Eng Data 2007;52:1552–6.

[40] Nagarwal RC, Kumar R, Dhanawat M, Das N, Pandit JK. Nanocrystal technology inthe delivery of poorly soluble drugs: an overview. Curr DrugDeliv 2011;8:398–406.

[41] Hartel RW. Crystallization in foods. Gaithersburg, MD: Aspen Publishers; 2001.[42] Dickinson E, Goller MI, McClements DJ, Peasgood S, Povey MJW. Ultrasonic mon-

itoring of crystallization in an oil-in-water emulsion. J Chem Soc Faraday Trans1990;86:1147–8.

[43] Gulseren I, Coupland JN. The effect of emulsifier type and droplet size on phasetransitions in emulsified even-numbered n-alkanes. J Am Oil Chem Soc2007;84:621–9.

[44] Gunther E, Schmid T, Mehling H, Hiebler S, Huang L. Subcooling in hexadecaneemulsions. Int J Refrig Rev Int Froid 2010;33:1605–11.

[45] McClements DJ, Dungan SR, German JB, Simoneau C, Kinsella JE. Droplet size andemulsifier type affect crystallization and melting of hydrocarbon-in-water emul-sions. J Food Sci 1993;58:1148-&.

[46] Kashchiev D, van Rosmalen GM. Review: nucleation in solutions revisited. CrystRes Technol 2003;38:555–74.

[47] Maris HJ. Introduction to the physics of nucleation. C R Phys 2006;7:946–58.[48] Leubner IH. Particle nucleation and growth models. Curr Opin Colloid Interface

Sci 2000;5:151–9.[49] Herhold AB, Ertas D, Levine AJ, King HE. Impurity mediated nucleation in

hexadecane-in-water emulsions. Phys Rev E 1999;59:6946–55.[50] Kashchiev D, Borissova A, Hammond RB, Roberts KJ. Effect of cooling rate on the

critical undercooling for crystallization. J Cryst Growth 2010;312:698–704.[51] Turnbull D, Fisher JC. Rate of nucleation in condensed systems. J Chem Phys

1949;17:71–3.[52] Balibar S, Caupin F. Nucleation of crystals from their liquid phase. C R Phys

2006;7:988–99.[53] Carlert S, Palsson A, Hanisch G, von Corswant C, Nilsson C, Lindfors L, et al. Pre-

dicting intestinal precipitation—a case example for a basic BCS class II drug.Pharm Res 2010;27:2119–30.

[54] Sugano K. Introduction to computational oral absorption simulation. Expert OpinDrug Metab Toxicol 2009;5:259–93.

[55] Erdemir D, Lee AY, Myerson AS. Nucleation of crystals from solution: classicaland two-step models. Acc Chem Res 2009;42:621–9.

[56] Chen J, Sarma B, Evans JMB, Myerson AS. Pharmaceutical crystallization. CrystGrowth Des 2011;11:887–95.

[57] Vekilov PG. Nucleation of crystals in solution. In: Wang M, Tsukamoto K, Wu D,editors. Selected topics on crystal growth, vol. 1270; 2010. p. 60–77.

[58] Vekilov PG. The two-step mechanism of nucleation of crystals in solution. Nano-scale 2010;2:2346–57.

[59] Van Eerdenbrugh B, Vermant J, Martens JA, Froyen L, Van Humbeeck J, Van denMonter G, et al. Solubility increases associated with crystalline drug nanoparti-cles: methodologies and significance. Mol Pharm 2010;7:1858–70.

[60] Warren DB, Benameur H, Porter CJH, Pouton CW. Using polymeric precipitationinhibitors to improve the absorption of poorly water-soluble drugs: a mechanis-tic basis for utility. J Drug Target 2010;18:704–31.

[61] Bevernage J, Forier T, Brouwers J, Tack J, Annaert P, Augustijns P. Excipient-me-diated supersaturation stabilization in human intestinal fluids. Mol Pharm2011;8:564–70.

[62] Brouwers J, Brewster ME, Augustijns P. Supersaturating drug delivery systems:the answer to solubility-limited oral bioavailability? J Pharm Sci 2009;98:2549–72.


[63] Herrera ML, Hartel RW. Effect of processing conditions on crystallization kineticsof a milk fat model system. J Am Oil Chem Soc 2000;77:1177–87.

[64] Arnold YE, Imanidis G, Kuentz MT. Advancing in-vitro drug precipitation testing:new process monitoring tools and a kinetic nucleation and growth model.J Pharm Pharmacol 2011;63:333–41.

[65] Lee AY, Myerson AS. Particle engineering: fundamentals of particle formationand crystal growth. MRS Bull 2006;31:881–6.

[66] Timms RE. Crystallization of fats. Chem Ind 1991:342.[67] Li JL, Liu XY. Microengineering of soft functional materials by controlling the

fiber network formation. J Phys Chem B 2009;113:15467–72.[68] Li JL, Liu XY. Architecture of supramolecular soft functional materials: from under-

standing to micro-/nanoscale engineering. Adv Funct Mater 2010;20:3196–216.[69] Li JL, Yuan B, Liu XY, Xu HY. Microengineering of supramolecular soft materials

by design of the crystalline fiber networks. Cryst Growth Des 2010;10:2699–706.[70] Duffy N, Blonk HCG, Beindorff CM, Cazade M, Bot A, Duchateau G. Organogel-

based emulsion systems, micro-structural features and impact on in vitro diges-tion. J Am Oil Chem Soc 2009;86:733–41.

[71] Bot A, den Adel R, Regkos C, Sawalha H, Venema P, Floter E. Structuring in beta-sitosterol plus gamma-oryzanol-based emulsion gels during various stages of atemperature cycle. Food Hydrocolloids 2011;25:639–46.

[72] Caira MR. Crystalline polymorphism of organic compounds. Design of organicsolidsBerlin: Springer-Verlag; 1998. p. 163–208. 33.

[73] Himawan C, Starov VM, Stapley AGF. Thermodynamic and kinetic aspects of fatcrystallization. Adv Colloid Interface Sci 2006;122:3–33.

[74] Sato K. Crystallization behaviour of fats and lipids—a review. Chem Eng Sci2001;56:2255–65.

[75] Davey RJ, Blagden N, Potts GD, Docherty R. Polymorphism in molecular crystals:stabilization of a metastable form by conformational mimicry. J Am Chem Soc1997;119:1767–72.

[76] Bunjes H, Koch MHJ. Saturated phospholipids promote crystallization but slowdown polymorphic transitions in triglyceride nanoparticles. J Control Release2005;107:229–43.

[77] Bunjes H. Structural properties of solid lipid based colloidal drug delivery sys-tems. Curr Opin Colloid Interface Sci 2011;16:405–11.

[78] Coupland JN. Crystallization in emulsions. Curr Opin Colloid Interface Sci 2002;7:445–50.

[79] Kozisek Z, Sato K, Ueno S, Demo P. Formation of crystal nuclei near critical super-saturation in small volumes. J Chem Phys 2011;134.

[80] Montenegro R, Antonietti M, Mastai Y, Landfester K. Crystallization in miniemul-sion droplets. J Phys Chem B 2003;107:5088–94.

[81] Turnbull D, Cormia RL. Kinetics of crystal nucleation in some normal alkane liq-uids. J Chem Phys 1961;34:820-&.

[82] Gulseren I, Coupland JN. Surface melting in alkane emulsion droplets as affectedby surfactant type. J Am Oil Chem Soc 2008;85:413–9.

[83] Palanuwech J, Coupland JN. Effect of surfactant type on the stability of oil-in-wateremulsions to dispersed phase crystallization. Colloids Surf A Physicochem EngAspects 2003;223:251–62.

[84] Arima S, Ueno S, Ogawa A, Sato K. Scanning microbeam small-angle X-ray dif-fraction study of interfacial heterogeneous crystallization of fat crystals in oil-in-water emulsion droplets. Langmuir 2009;25:9777–84.

[85] Awad T, Sato K. Effects of hydrophobic emulsifier additives on crystallization be-havior of palm mid fraction in oil-in-water emulsion. J Am Oil Chem Soc2001;78:837–42.

[86] Awad T, Sato K. Acceleration of crystallisation of palm kernel oil in oil-in-wateremulsion by hydrophobic emulsifier additives. Colloids Surf B Biointerfaces2002;25:45–53.

[87] Awad TS. Ultrasonic studies of the crystallization behavior of two palm fats O/Wemulsions and its modification. Food Res Int 2004;37:579–86.

[88] Katsuragi T, Kaneko N, Sato K. Effects of addition of hydrophobic sucrose fattyacid oligoesters on crystallization rates of n-hexadecane in oil-in-water emul-sions. Colloids Surf B Biointerfaces 2001;20:229–37.

[89] McClements DJ, Dickinson E, Dungan SR, Kinsella JE, Ma JG, Povey MJW. Effect ofemulsifier type on the crystallization kinetics of oil-in-water emulsions contain-ing a mixture of solid and liquid droplets. J Colloid Int Sci 1993;160:293–7.

[90] Feltham DL, Garside J. A mathematical model of crystallization in an emulsion.J Chem Phys 2005;122.

[91] Boode K, Bisperink C, Walstra P. Destabilization of O/W emulsions containing fatcrystals by temperature cycling. Colloids Surf 1991;61:55–74.

[92] Boode K, Walstra P. Partial coalescence in oil-in-water emulsions .1. Nature ofthe aggregation. Colloids Surf A Physicochem Eng Aspects 1993;81:121–37.

[93] Helgason T, Awad TS, Kristbergsson K, McClements DJ, Weiss J. Influence of poly-morphic transformations on gelation of tripalmitin solid lipid nanoparticle sus-pensions. J Am Oil Chem Soc 2008;85:501–11.

[94] Hauss DJ, Fogal SE, Ficorilli JV, Price CA, Roy T, Jayara AA, et al. Lipid-based deliv-ery systems for improving the bioavailability and lymphatic transport of a poorlywater-soluble LTB4 inhibitor. J Pharm Sci 1998;87:164–9.

[95] Golding M, Wooster TJ, Day L, Xu M, Lundin L, Keogh J, et al. Impact of gastricstructuring on the lipolysis of emulsified lipids. Soft Matter 2011;7:3513–23.

[96] Kesisoglou F, Panmai S, Wu YH. Nanosizing—oral formulation development andbiopharmaceutical evaluation. Adv Drug Deliv Rev 2007;59:631–44.

[97] Joshi MD, Muller RH. Lipid nanoparticles for parenteral delivery of actives. Eur JPharm Biopharm 2009;71:161–72.

[98] Bunjes H, Koch MH, Westesen K. Influence of emulsifiers on the crystallization ofsolid lipid nanoparticles. J Pharm Sci 2003;92:1509–20.

[99] Bunjes H, Koch MHJ, Westesen K. Effect of particle size on colloidal solid triglyc-erides. Langmuir 2000;16:5234–41.

[100] Jores K, Mehnert W, Drechsler M, Bunjes H, Johann C, Mader K. Investigations onthe structure of solid lipid nanoparticles (SLN) and oil-loaded solid lipid nano-particles by photon correlation spectroscopy, field-flow fractionation and trans-mission electron microscopy. J Control Release 2004;95:217–27.

[101] Bunjes H, Steiniger F, Richter W. Visualizing the structure of triglyceride nano-particles in different crystal modifications. Langmuir 2007;23:4005–11.

[102] Uskokovic V. Morphological study of emulsion-assisted cholesterol precipitationprocesses. Mol Cryst Liq Cryst 2007;474:77–88.

[103] Liu Y, Kathan K, Saad W, Prud'homme RK. Ostwald ripening of beta-carotenenanoparticles. Phys Rev Lett 2007;98.

[104] Adam-Berret M, Boulard M, Riaublanc A, Mariette F. Evolution of fat crystal net-work microstructure followed by NMR. J Agric Food Chem 2011;59:1767–73.

[105] Awad TS, Helgason T, Kristbergsson K, Decker EA, Weiss J, McClements DJ. Effectof cooling and heating rates on polymorphic transformations and gelation of tri-palmitin solid lipid nanoparticle (SLN) suspensions. Food Biophys 2008;3:155–62.

[106] Westesen K, Bunjes H, Koch MHJ. Physicochemical characterization of lipidnanoparticles and evaluation of their drug loading capacity and sustained re-lease potential. J Control Release 1997;48:223–36.

[107] Helgason T, Awad TS, Decker E, Kristbergsson K, McClements DJ, Weiss J. Influ-ence of polymorphic transformations on gelation of tripalmitin solid lipid nano-particle (SLN) suspensions. submitted to Journal of the American Oil ChemistsSociety 2007.

[108] Helgason T, Awad TS, Kristbergsson K, McClements DJ, Weiss J. Effect of surfac-tant surface coverage on formation of solid lipid nanoparticles (SLN). J ColloidInterface Sci 2009;334:75–81.

[109] Arima S, Ueji T, Ueno S, Ogawa A, Sato K. Retardation of crystallization-induceddestabilization of PMF-in-water emulsion with emulsifier additives. ColloidsSurf B Biointerfaces 2007;55:98–106.

[110] Israelachvili J. Intermolecular and surface forces. Second edition. London, UK:Academic Press; 1992.

[111] Fredrick E, Walstra P, Dewettinck K. Factors governing partial coalescence in oil-in-water emulsions. Adv Colloid Interface Sci 2010;153:30–42.

[112] Freitas C, Muller RH. Correlation between long-term stability of solid lipid nano-particles (SLN (TM)) and crystallinity of the lipid phase. Eur J Pharm Biopharm1999;47:125–32.

[113] Anton N, Benoit JP, Saulnier P. Design and production of nanoparticles formulat-ed from nano-emulsion templates—a review. J Control Release 2008;128:185–99.

[114] Anton N, Vandamme TF. The universality of low-energy nano-emulsification. IntJ Pharm 2009;377:142–7.

[115] Leong TSH, Wooster TJ, Kentish SE, Ashokkumar M. Minimising oil droplet sizeusing ultrasonic emulsification. Ultrason Sonochem 2009;16:721–7.

[116] Acosta E. Bioavailability of nanoparticles in nutrient and nutraceutical delivery.Curr Opin Colloid Interface Sci 2009;14:3–15.

[117] Wooster TJ, Golding M, Sanguansri P. Impact of oil type on nanoemulsion forma-tion and Ostwald ripening stability. Langmuir 2008;24:12758–65.

[118] Gutierrez JM, Gonzalez C, Maestro A, Sole I, Pey CM, Nolla J. Nano-emulsions:new applications and optimization of their preparation. Curr Opin Colloid Inter-face Sci 2008;13:245–51.

[119] Freitas S, Merkle HP, Gander B. Microencapsulation by solvent extraction/evaporation: reviewing the state of the art of microsphere preparation pro-cess technology. J Control Release 2005;102:313–32.

[120] Bouchemal K, Briancon S, Perrier E, Fessi H. Nano-emulsion formulation usingspontaneous emulsification: solvent, oil and surfactant optimisation. Int JPharm 2004;280:241–51.

[121] Troncoso E, Aguilera JM, McClements DJ. Fabrication, characterization and lipasedigestibility of food-grade nanoemulsions. Food Hydrocolloids 2012;27:355–63.

[122] Chu BS, Ichikawa S, Kanafusa S, Nakajima M. Preparation and characterization ofbeta-carotene nanodispersions prepared by solvent displacement technique.J Agric Food Chem 2007;55:6754–60.

[123] McClements DJ. Nanoemulsions versus microemulsions: terminology, differ-ences, and similarities. Soft Matter 2012;8:1719–29.

[124] Thanatuksorn P, Kawai K, Hayakawa M, Hayashi M, Kajiwara K. Improvement ofthe oral bioavailability of coenzyme Q(10) by emulsification with fats and emul-sifiers used in the food industry. LWT—Food Sci Technol 2009;42:385–90.

[125] Kommuru TR, Gurley B, Khan MA, Reddy IK. Self-emulsifying drug delivery sys-tems (SEDDS) of coenzyme Q(10): formulation development and bioavailabilityassessment. Int J Pharm 2001;212:233–46.

[126] Walstra P. Principles of emulsion formation. Chem Eng Sci 1993;48:333.[127] Jafari SM, Assadpoor E, He YH, Bhandari B. Re-coalescence of emulsion droplets

during high-energy emulsification. Food Hydrocolloids 2008;22:1191–202.[128] Kentish S, Wooster TJ, Ashokkumar A, Balachandran S, Mawson R, Simons L. The

use of ultrasonics for nanoemulsion preparation. 3rd Innovative Foods CentreConference on Food Innovation: Emerging Science, Technologies and Applica-tions. Melbourne, Australia: Elsevier Sci Ltd; 2006. p. 170–5.

[129] Salager JL, Forgiarini A, Marquez L, Pena A, Pizzino A, Rodriguez MP, et al. Usingemulsion inversion in industrial processes. Adv Colloid Interface Sci 2004;108:259–72.

[130] Queste S, Salager JL, Strey R, Aubry JM. The EACN scale for oil classification revis-ited thanks to fish diagrams. J Colloid Interface Sci 2007;312:98–107.

[131] Leal-Calderon F, Schmitt V, Bibette J. Emulsion science: basic principles. SpringerVerlag; 2007.

[132] Salager JL, Marquez L, Mira I, Pena A, Tyrode E, Zambrano NB. Principles of emul-sion formulation engineering. In: Mittal KL, Shah DO, editors. Adsorption and ag-gregation of surfactants in solution., vol. 109; 2003. p. 501–23.


[133] Solans C, Esquena J, Forgiarini AM, Uson N, Morales D, Izquierdo P, et al. Nano-emulsions: formation, properties, and applications. In: Mittal KL, Shah DO, edi-tors. Adsorption and aggregation of surfactants in solution., vol. 109; 2003.p. 525–54.

[134] Choi SJ, Decker EA, Henson L, Popplewell LM, McClements DJ. Stability of citral inoil-in-water emulsions prepared with medium-chain triacylglycerols and triace-tin. J Agric Food Chem 2009;57:11349–53.

[135] Harms M, Muller-Goymann CC. Solid lipid nanoparticles for drug delivery. J DrugDeliv Sci Technol 2011;21:89–99.

[136] Pardeike J, Hommoss A, Muller RH. Lipid nanoparticles (SLN, NLC) in cosmeticand pharmaceutical dermal products. Int J Pharm 2009;366:170–84.

[137] Ghosh S, Peterson DG, Coupland JN. Effects of droplet crystallization and meltingon the aroma release properties of a model oil-in-water emulsion. J Agric FoodChem 2006;54:1829–37.

[138] Ghosh S, Peterson DG, Coupland JN. Aroma release from solid droplet emulsions:effect of lipid type. J Am Oil Chem Soc 2007;84:1001–14.

[139] Mei LY, Choi SJ, Alamed J, Henson L, Popplewell M, McClements DJ, et al. Citralstability in oil-in-water emulsions with solid or liquid octadecane. J Agric FoodChem 2010;58:533–6.

[140] Okuda S, McClements DJ, Decker EA. Impact of lipid physical state on the oxida-tion of methyl linolenate in oil-in-water emulsions. J Agric Food Chem 2005;53:9624–8.

[141] Boon CS, McClements DJ, Weiss J, Decker EA. Factors influencing the chemicalstability of carotenoids in foods. Crit Rev Food Sci Nutr 2010;50:515–32.

[142] McClements DJ, Decker EA. Lipid oxidation in oil-in-water emulsions: impactof molecular environment on chemical reactions in heterogeneous food systems.J Food Sci 2000;65:1270–82.

[143] McClements DJ, Decker EA, Park Y, Weiss J. Structural design principles for deliv-ery of bioactive components in nutraceuticals and functional foods. Crit RevFood Sci Nutr 2009;49:577–606.

[144] Failla ML, Huo T, Thakkar SK. In vitro screening of relative bioaccessibility of ca-rotenoids from foods. 10th Asian Congress of Nutrition. Taipei, Taiwan: H E CPress, Healthy Eating Club Pty Ltd; 2007. p. 200–3.

[145] Higuera-Ciapara I, Felix-Valenzuela L, Goycoolea FM. Astaxanthin: a review of itschemistry and applications. Crit Rev Food Sci Nutr 2006;46:185–96.

[146] Singh P, Goyal GK. Dietary lycopene: its properties and anticarcinogenic effects.Compr Rev Food Sci Food Saf 2008;7:255–70.

[147] Abdel-Aal ESM, Akhtar MH. Recent advances in the analyses of carotenoids andtheir role in human health. Curr Pharmaceut Anal 2006;2:195–204.

[148] Boon CS, McClements DJ, Weiss J, Decker EA. Role of iron and hydroperoxides inthe degradation of lycopene in oil-in-water emulsions. J Agric Food Chem2009;57:2993–8.

[149] Helgason T, Awad TS, Kristbergsson K, Decker EA, McClements DJ, Weiss J.Impact of surfactant properties on oxidative stability of beta-carotene en-capsulated within solid lipid nanoparticles. J Agric Food Chem 2009;57:8033–40.

[150] Cornacchia L, Roos YH. State of dispersed lipid carrier and interface compositionas determinants of beta-carotene stability in oil-in-water emulsions. J Food Sci2011;76:C1211–8.

[151] Lundin L, Golding M. Structure design for healthy food. Aust J Dairy Technol2009;64:68–74.

[152] McClements DJ, Li Y. Structured emulsion-based delivery systems: controllingthe digestion and release of lipophilic food components. Adv Colloid InterfaceSci 2010;159:213–28.

[153] Singh H, Ye AQ, Horne D. Structuring food emulsions in the gastrointestinal tractto modify lipid digestion. Prog Lipid Res 2009;48:92–100.

[154] Bouwmeester H, Dekkers S, NoordamMY, Hagens WI, Bulder AS, de Heer C, et al.Review of health safety aspects of nanotechnologies in food production. RegulToxicol Pharmacol 2009;53:52–62.

[155] Hagens WI, Oomen AG, de Jong WH, Cassee FR, Sips A. What do we (need to)know about the kinetic properties of nanoparticles in the body? Regul ToxicolPharmacol 2007;49:217–29.

[156] Hu L, Mao ZW, Gao CY. Colloidal particles for cellular uptake and delivery.J Mater Chem 2009;19:3108–15.

[157] McClements DJ, Decker EA, Park Y. Controlling lipid bioavailability throughphysicochemical and structural approaches. Crit Rev Food Sci Nutr 2009;49:48–67.

[158] van Aken GA. Relating food emulsion structure and composition to the way it isprocessed in the gastrointestinal tract and physiological responses: what are theopportunities? Food Biophys 2010;5:258–83.

[159] van Aken GA. Modelling texture perception by soft epithelial surfaces. Soft Mat-ter 2010;6:826–34.

[160] Kong F, Singh RP. Disintegration of solid foods in human stomach. J Food Sci2008;73:R67–80.

[161] Ferrua MJ, Singh RP. Modeling the fluid dynamics in a human stomach to gaininsight of food digestion. J Food Sci 2010;75:R151–62.

[162] Golding M, Wooster TJ. The influence of emulsion structure and stability on lipiddigestion. Curr Opin Colloid Interface Sci 2010;15:90–101.

[163] Lundin L, Golding M, Wooster TJ. Understanding food structure and function indeveloping food for appetite control. Nutr Diet 2008;65:S79–85.

[164] Marciani L, Faulks R, Wickham MSJ, Bush D, Pick B, Wright J, et al. Effect of intra-gastric acid stability of fat emulsions on gastric emptying, plasma lipid profileand postprandial satiety. Br J Nutr 2009;101:919–28.

[165] Marciani L, Wickham M, Singh G, Bush D, Pick B, Cox E, et al. Delaying gastricemptying and enhancing cholecystokinin release and satiety by using acid stablefat emulsions. Gastroenterology 2006;130:A227.

[166] Marciani L, Wickham M, Singh G, Bush D, Pick B, Cox E, et al. Enhancement ofintragastric acid stability of a fat emulsion meal delays gastric emptying and in-creases cholecystokinin release and gallbladder contraction. Am J Physiol Gas-trointest Liver Physiol 2007;292:G1607–13.

[167] Day JPR, Rago G, Domke KF, Velikov KP, Bonn M. Label-free imaging of lipophilicbioactive molecules during lipid digestion by multiplex coherent anti-stokesraman scattering microspectroscopy. J Am Chem Soc 2010;132:8433–9.

[168] Bonnaire L, Sandra S, Helgason T, Decker EA, Weiss J, McClements DJ. Influenceof lipid physical state on the in vitro digestibility of emulsified lipids. J AgricFood Chem 2008;56:3791–7.

[169] Nik AM, Wright AJ, Corredig M. Micellization of beta-carotene from soy-proteinstabilized oil-in-water emulsions under in vitro conditions of lipolysis. J Am OilChem Soc 2011;99:1397–407.

[170] McClements DJ, Decker EA, Park Y. Physicochemical and structural aspects oflipid digestion. In: McClements DJ, editor. Understanding and controlling the mi-crostructure of complex foods. Boca Raton, FL: CRC Press; 2007. p. 483–503.

[171] Johnson LR. Gastrointestinal physiology. 6th edition. St. Louis, MI: Mosby; 2001.[172] De JongWH, HagensWI, Krystek P, Burger MC, Sips A, Geertsma RE. Particle size-

dependent organ distribution of gold nanoparticles after intravenous adminis-tration. Biomaterials 2008;29:1912–9.

[173] des Rieux A, Fievez V, Garinot M, Schneider YJ, Preat V. Nanoparticles as poten-tial oral delivery systems of proteins and vaccines: a mechanistic approach.J Control Release 2006;116:1–27.

[174] Sahay G, Alakhova DY, Kabanov AV. Endocytosis of nanomedicines. J Control Re-lease 2010;145:182–95.

[175] Sahay G, Kim JO, Kabanov AV, Bronich TK. The exploitation of differential endo-cytic pathways in normal and tumor cells in the selective targeting of nanopar-ticulate chemotherapeutic agents. Biomaterials 2010;31:923–33.

[176] Yuan H, Chen J, Du YZ, Hu FQ, Zeng S, Zhao HL. Studies on oral absorption of stea-ric acid SLN by a novel fluorometric method. Colloids Surf B Biointerfaces2007;58:157–64.

[177] Unfried K, Albrecht C, Klotz LO, Von Mikecz A, Grether-Beck S, Schins RPF. Cellu-lar responses to nanoparticles: target structures and mechanisms. Nanotoxicol-ogy 2007;1:52–71.

[178] ArmandM, Borel P, Ythier P, Dutot G,Melin C, SenftM, et al. Effects of droplet size,triacylglycerol composition, and calcium on the hydrolysis of complex emulsionsby pancreatic lipase—an invitro study. J Nutr Biochem 1992;3:333–41.

[179] Lee SJ, Choi SJ, Li Y, Decker EA, McClements DJ. Protein-stabilized nanoemulsionsand emulsions: comparison of physicochemical stability, lipid oxidation, and li-pase digestibility. J Agric Food Chem 2011;59:415–27.

[180] Torcello-Gomez A, Maldonado-Valderrama J, Martin-Rodriguez A, McClements DJ.Physicochemical properties and digestibility of emulsified lipids in simulated intes-tinal fluids: influence of interfacial characteristics. Soft Matter 2011;7:6167–77.

[181] Mun S, Decker EA, Park Y, Weiss J, McClements DJ. Influence of interfacial com-position on in vitro digestibility of emulsified lipids: potential mechanism forchitosan's ability to inhibit fat digestion. Food Biophys 2006;1:21–9.

[182] Chu BS, Rich GT, Ridout MJ, Faulks RM, WickhamMSJ, Wilde PJ. Modulating pan-creatic lipase activity with galactolipids: effects of emulsion interfacial composi-tion. Langmuir 2009;25:9352–60.

[183] Li Y, McClements DJ. New mathematical model for interpreting pH-Stat diges-tion profiles: impact of lipid droplet characteristics on in vitro digestibility.J Agric Food Chem 2010;58:8085–92.

[184] Fernandez S, Chevrier S, Ritter N, Mahler B, Demarne F, Carriere F, et al. In vitrogastrointestinal lipolysis of four formulations of piroxicam and cinnarizine withthe self emulsifying excipients labrasol (R) and gelucire (R) 44/14. Pharm Res2009;26:1901–10.

[185] Colle IJP, Lemmens L, Tolesa GN, Van Buggenhout S, De Vleeschouwer K, VanLoey AM, et al. Lycopene degradation and isomerization kinetics during thermalprocessing of an olive oil/tomato emulsion. J Agric Food Chem 2010;58:12784–9.

[186] Bou R, Boon C, Kweku A, Hidalgo D, Decker EA. Effect of different antioxidants onlycopene degradation in oil-in-water emulsions. Eur J Lipid Sci Technol 2011;113:724–9.

[187] Boon CS, Xu Z, Yue X, McClements DJ, Weiss J, Decker EA. Factors affecting lyco-pene oxidation in oil-in-water emulsions. J Agric Food Chem 2008;56:1408–14.

[188] Anarjan N, Mirhosseini H, Baharin BS, Tan CP. Effect of processing conditions onphysicochemical properties of sodium caseinate-stabilized astaxanthin nanodis-persions. LWT—Food Sci Technol 2011;44:1658–65.

[189] Ribeiro HS, Rico LG, Badolato GG, Schubert H. Production of O/W emulsions con-taining astaxanthin by repeated Premix membrane emulsification. J Food Sci2005;70:E117–23.

[190] Hou ZQ, Gao YX, Yuan F, Liu YW, Li CL, Xu DX. Investigation into the physico-chemical stability and rheological properties of beta-carotene emulsion stabi-lized by soybean soluble polysaccharides and chitosan. J Agric Food Chem2010;58:8604–11.

[191] Mao LK, Yang J, XuDX, Yuan F, Gao YX. Effects of homogenizationmodels and emul-sifiers on the physicochemical properties of carotene nanoemulsions. J Dispers SciTechnol 2010;31:986–93.

[192] Yin LJ, Chu BS, Kobayashi I, Nakajima M. Performance of selected emulsifiers andtheir combinations in the preparation of beta-carotene nanodispersions. FoodHydrocolloids 2009;23:1617–22.

[193] Neves MA, Ribeiro HS, Kobayashi I, Nakajima M. Encapsulation of lipophilicbioactive molecules by microchannel emulsification. Food Biophys 2008;3:126–31.

[194] Chu BS, Ichikawa S, Kanafusa S, Nakajima M. Preparation of protein-stabilizedbeta-carotene nanodispersions by emulsification–evaporation method. J AmOil Chem Soc 2007;84:1053–62.


[195] Yuan Y, Gao YX, Zhao J, Mao L. Characterization and stability evaluation of beta-carotene nanoemulsions prepared by high pressure homogenization under var-ious emulsifying conditions. Food Res Int 2008;41:61–8.

[196] Silva HD, Cerqueira MA, Souza BWS, Ribeiro C, Avides MC, Quintas MAC, et al.Nanoemulsions of beta-carotene using a high-energy emulsification–evaporationtechnique. J Food Eng 2011;102:130–5.

[197] Sato Y, Kobayashi M, Itagaki S, Hirano T, Noda T, Mizuno S, et al. Pharmacokineticproperties of lutein emulsion after oral administration to rats and effect of foodintake on plasma concentration of lutein. Biopharm Drug Dispos 2011;32:151–8.

[198] Yoo JH, Shanmugam S, Thapa P, Lee ES, Balakrishnan P, Baskaran R, et al. Novelself-nanoemulsifying drug delivery system for enhanced solubility and dissolu-tion of lutein. Arch Pharm Res 2010;33:417–26.

[199] Losso JN, Khachatryan A, Ogawa M, Godber JS, Shih F. Random centroid optimi-zation of phosphatidylglycerol stabilized lutein-enriched oil-in-water emulsionsat acidic pH. Food Chem 2005;92:737–44.

[200] Tyssandier V, Lyan B, Borel P. Main factors governing the transfer of carotenoidsfrom emulsion lipid droplets to micelles. Biochim Biophys Acta Mol Cell BiolLipids 2001;1533:285–92.

[201] Wang XY, Jiang Y, Wang YW, Huang MT, Ho CT, Huang QR. Enhancing anti-inflammation activity of curcumin through O/W nanoemulsions. Food Chem2008;108:419–24.

[202] Nayak AP, Tiyaboonchai W, Patankar S, Madhusudhan B, Souto EB. Curcuminoids-loaded lipid nanoparticles: novel approach towards malaria treatment. ColloidsSurf B Biointerfaces 2010;81:263–73.

[203] Choi CH, Kim SH, Shanmugam S, Baskaran R, Park JS, Yong CS, et al. Relative bio-availability of coenzyme Q10 in emulsion and liposome formulations. BiomolTher 2010;18:99–105.

[204] Yue Y, Zhou HF, Liu GL, Li Y, Yan ZM, Duan MX. The advantages of a novel CoQ10delivery system in skin photo-protection. Int J Pharm 2010;392:57–63.

[205] Augustin MA, Abeywardena MY, Patten G, Head R, Lockett T, De Luca A, et al. Ef-fects of microencapsulation on the gastrointestinal transit and tissue distribu-tion of a bioactive mixture of fish oil, tributyrin and resveratrol. J Funct Foods2011;3:25–37.

[206] Hung CF, Chen JK, Liao MH, Lo HM, Fang JY. Development and evaluation ofemulsion-liposome blends for resveratrol delivery. J Nanosci Nanotechnol2006;6:2950–8.

[207] Hung CF, Fang CL, Liao MH, Fang JY. The effect of oil components on the physico-chemical properties and drug delivery of emulsions: Tocol emulsion versus lipidemulsion. Int J Pharm 2007;335:193–202.

[208] Leong WF, Cheong KW, Lai OM, Long K, Man YBC, Misran M, et al. Response sur-face modeling of processing parameters for the preparation of phytosterol nano-dispersions using an emulsification–evaporation technique. J Am Oil Chem Soc2011;88:717–25.

[209] Cercaci L, Rodriguez-Estrada MT, Lercker G, Decker EA. Phytosterol oxidation inoil-in-water emulsions and bulk oil. Food Chem 2007;102:161–7.

[210] Salentinig S, Sagalowicz L, Leser ME, Tedeschi C, Glatter O. Transitions in the in-ternal structure of lipid droplets during fat digestion. Soft Matter 2011;7:650–61.

Crystals and crystallization in oil-in-water emulsions: Implications for emulsion-based delivery systems

Documents