at SciVerse ScienceDirect

Food Hydrocolloids 29 (2012) 398e406

Contents lists available

Food Hydrocolloids

journal homepage: www.elsevier .com/locate/ foodhyd

Fabrication of ultrafine edible emulsions: Comparison of high-energyand low-energy homogenization methods

Ying Yang, Christopher Marshall-Breton, Martin E. Leser, Alexander A. Sher, David Julian McClements*

Biopolymers and Colloids Laboratory, Department of Food Science, University of Massachusetts, Amherst, MA 01003, USA

a r t i c l e i n f o

Article history:Received 22 December 2011Accepted 11 April 2012

Keywords:Spontaneous emulsificationMicrofluidizerComparisonTween 80Tween 85

* Corresponding author. Tel.: þ1 (413) 545 1019; faE-mail address: mcclements@foodsci.umass.edu (D

0268-005X/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.foodhyd.2012.04.009

a b s t r a c t

Emulsions containing ultrafine droplets (r < 100 nm) have a number of potential advantages overconventional emulsions for the encapsulation and delivery of lipophilic substances in foods and bever-ages: high optical clarity; high physical stability; increased bioavailability. These ultrafine emulsions canbe fabricated from high-energy or low-energy homogenization methods, which each have advantagesand limitations. In this study, we compared a high-energy method (microfluidization) with a low-energymethod (spontaneous emulsification) for forming oil-in-water emulsions from food-grade ingredients(medium chain triglycerides and Tweens). The influence of surfactant type (Tween 80, Tween 85, andTween 80/Tween 85) and surfactant-to-oil ratio (SOR ¼ 0.1e5) on the formation of emulsions wasexamined. Both the low- and high-energy methods were able to produce emulsions with ultrafinedroplets (r < 100 nm). The microfluidization method required high-energy inputs and dedicatedequipment, but could produce ultrafine emulsions at much lower surfactant-to-oil ratio (SOR < 0.1). Onthe other hand, the spontaneous emulsification method only required simple mixing, but it needed muchhigher surfactant-to-oil ratios (SOR > 0.5) to produce droplets with r < 100 nm. This study has importantimplications for the development of food-grade delivery systems to encapsulate lipophilic substances,such as flavors, colors, vitamins, and nutraceuticals.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

There is considerable interest within the food and beverageindustries in the utilization of ultrafine emulsions (sometimesreferred to as nanoemulsions) to encapsulate and deliver lipophilicfunctional agents, such as flavors, colors, antimicrobials, micro-nutrients, and nutraceuticals (McClements, 2010; McClements,Decker, & Weiss, 2007; Sagalowicz & Leser, 2010; Sanguansri &Augustin, 2006; Velikov & Pelan, 2008). Ultrafine emulsions canbe fabricated from food-grade ingredients using relatively simpleprocessing operations, such as mixing, shearing, and homogeni-zation (McClements, 2011; McClements & Rao, 2011). Ultrafineemulsions are thermodynamically unstable systems that typicallyconsist of surfactant, oil, and water (Mason, Wilking, Meleson,Chang, & Graves, 2006; Sonneville-Aubrun, Simonnet, & L’Alloret,2004; Tadros, Izquierdo, Esquena, & Solans, 2004). Oil-in-wateremulsions contain surfactant-coated lipid droplets dispersedwithin an aqueous continuous phase. The radius of the droplets inultrafine emulsions (r< 100 nm) is relatively small compared to the

x: þ1 (413) 545 1262..J. McClements).

All rights reserved.

wavelength of light (l¼ 390e750 nm) and so they tend to be eithertransparent or only slightly turbid (McClements, 2002; Velikov andPelan, 2008). One of the most promising potential applications ofultrafine emulsions is therefore to incorporate lipophilic activeingredients (e.g., vitamins, nutraceuticals, and antimicrobials) intoaqueous-based foods or beverages that need to remain opticallytransparent, such as some fortified waters, soft drinks, sauces, anddips (Velikov and Pelan, 2008). Another potential advantage ofthese systems is that they often have better stability to particleaggregation and gravitational separation than conventional emul-sions due to their small particle size (Tadros et al., 2004). Inparticular, Brownian motion effects become important for smallparticles (r < 100 nm), which counterbalances the tendency forgravitational separation to occur (McClements and Rao, 2011). Thefact that the interfacial layer surrounding very small oil dropletsmakes up an appreciable part of the total particle volume meansthat ultrafine emulsions can be produced that are much moreviscous than conventional emulsions with the same oil content(Mason et al., 2006; Sonneville-Aubrun et al., 2004; Tadros et al.,2004), which may have interesting applications in food productswhere a paste or gel is required. The small size of the droplets inultrafine emulsions has also been shown to increase the bioavail-ability of certain types of lipophilic substances (Acosta, 2009),

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406 399

which may be useful for increasing the bioactivity of somenutraceuticals.

In the food industry, ultrafine emulsions are usually producedusing high-energy methods, such as microfluidization, high pres-sure valve homogenization, and sonication (Einhorn-Stoll, Weiss, &Kunzek, 2002; Henry, Fryer, Frith, & Norton, 2009; Henry, Fryer,Frith, & Norton, 2010; Mason et al., 2006; Perrier-Cornet, Marie, &Gervais, 2005). These methods generate intense disruptive forcesthat mechanically breakup the oil phase into tiny droplets that aredispersedwithin the aqueous phase (Kentish et al., 2008). There area number of limitations in using high-energy methods to produceultrafine emulsions, such as high initial equipment and operatingcosts, high power requirements, potential for equipment break-down, and difficulties in producing very fine droplets from certainkinds of food ingredients (e.g., highly viscous oils or slowly-adsorbing emulsifiers) (McClements and Rao, 2011).

Ultrafine emulsions can also be produced using a variety of low-energy methods, such as phase inversion and spontaneous emul-sification methods (Anton, Benoit, & Saulnier, 2008; Anton &Vandamme, 2009; McClements and Rao, 2011). These low-energymethods are based on the spontaneous formation of fine oildroplets in surfactant-oil-water (SOW) systems under specificenvironmental conditions (composition, temperature, stirring),which are then trapped in a metastable state (Rao & McClements,2010; Sadurní, Solans, Azemar, & García-Celma, 2005). Low-energy approaches may have advantages over high-energyapproaches for certain applications within the food and beverageindustries: they are often more effective at producing very finedroplets; they have lower equipment and energy costs; they aresimpler to implement. On the other hand, there are also somepotential disadvantages of low-energy systems, including limita-tions on the types of oils and surfactants that can be used to formstable ultrafine emulsions, and the fact that relatively highsurfactant-to-oil ratios (SOR) are typically needed to produce them.

The purpose of this study was to compare the effectiveness ofa high-energymethod (microfluidization) that is widely used in the

Fig. 1. Underlying principles for droplet formation in high-energy (microfluidization)

food industry to produce ultrafine emulsions with a low-energymethod (spontaneous emulsification) that has been shown towork for many non-food applications. The microfluidizer works onthe principle of dividing an emulsion flowing through a channelinto two streams, passing each stream through a separate finechannel, and then directing the two streams at each other in aninteraction chamber. Intense disruptive forces are generated withinthe interaction chamber when the two fast moving streams ofemulsion impinge upon each other, leading to highly efficientdroplet disruption. A number of previous studies have shown thatmicrofluidization is particularly efficient at creating ultrafineemulsions from food-grade ingredients, and have identified someof the major factors that influence homogenization efficiency,including oilewater interfacial tension, oilewater viscosity ratio,and emulsifier type and concentration (Henry et al., 2010; Jafari, He,& Bhandari, 2007; Qian & McClements, 2011; Wooster, Golding, &Sanguansri, 2008). In the spontaneous emulsification method anemulsion is immediately formed when two liquids are mixedtogether (Fig. 1) (Anton & Vandamme, 2009; Miller, 1988; Pouton &Porter, 2006). Practically, this method can be carried out ina number of different ways to form ultrafine emulsions: thecompositions of the two phases can be varied; the environmentalconditions can be varied (e.g., temperature, pH and ionic strength);and/or, the mixing conditions can be varied (e.g., stirring speed, rateof addition, and order of addition) (Horn & Rieger, 2001). Thespontaneous emulsification method has been widely used toproduce ultrafine emulsions in the pharmaceutical industry, wherethey are usually referred to as self-nano-emulsifying drug deliverysystems (SNEDDS) (Nielsen et al., 2007; Nielsen, Petersen, &Mullertz, 2008).

In this study, we compared the efficacy of the microfluidizationand spontaneous emulsification methods at producing ultrafineemulsions from food-grade oils (MCT) and surfactants (Tweens),and we then tested the stability of the resulting emulsions toenvironmental stresses that they might encounter in the foodindustry. This research may be useful for identifying the most

and low-energy (spontaneous emulsification) methods of producing emulsions.

Fig. 2. Influence of surfactant concentration on the mean particle radius (r32) of 20 wt% MCT oil-in-water emulsions formed by a high-energy method (microfluidization).Emulsions were passed through the homogenizer three times at 10,000 psi.

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406400

appropriate homogenization method for different food andbeverage applications.

2. Materials and methods

2.1. Materials

Non-ionic surfactants, POE(20) sorbitan trioleate (Tween 85)and POE(20) sorbitan monoolieate (Tween 80), were purchasedfrom SigmaeAldrich Co. (St. Louis, MO). Medium chain triglyceride(MCT) oil (Miglyol 812) was purchased from Coletica (Northport,NY). The manufacturers reported that the fatty acid composition ofthe MCT was �2% caproic acid, 50e65% caprylic acid, 30e45%capric acid, �2% lauric acid and �1% myristic acid.

2.2. Emulsion preparation

2.2.1. Emulsions produced by microfluidizationOil, water and surfactant were mixed together in a beaker, and

then blended together using a high-shear mixer (Bamix, BiospecProducts, Bartlesville, OK) for 2 min to form a coarse emulsion. Thetotal weight of each sample was 100 g, and the overall compositionof the sampleswas 20wt%oil (MCT), 0.1 to 20wt% surfactant (Tween80 and/or Tween 85), and 79.9 to 60 wt% water (pH 3.5, buffersolution). The emulsion premix samples were then passed througha microfluidizer (M-110L, Microfluidics, Newton, MA) three times at10,000 psi (68.95 MPa) at ambient temperature using a specificinteraction chamber (F20Y 75 mm, Microfluidics, Newton, MA).

2.2.2. Emulsions produced by spontaneous emulsificationInitially, oil and surfactant were placed in a container and then

blended using a high-shear mixer for 2 min to prepare a homoge-neous solution. The oil-surfactant mixture was then poured intoa water phase (pH 3.5 buffer solution) at a rate of about 10 g perminute with continuous stirring (500 rpm) using a magnetic stirrerfor 15 min to keep the system homogeneous. At this pouring rate,the formation of an optically opaque emulsion was observed tooccur as soon as the oil-surfactant mixture came into contact withthe water phase. However, it was noticed that if the oil-surfactantmixture was poured into the water phase too quickly (>20 g perminute) then large clots were formed that were difficult todisperse, and therefore we always used slower rates.

2.3. Emulsion characterization

The particle size distributions of the samples were measuredusing a static light scattering instrument (Mastersizer 2000, Mal-vern Instruments, Malvern, U.K.). All measurements were per-formed on at least two freshly prepared samples (i.e., new sampleswere prepared for each series of experiments) andwere reported asmeans and standard deviations.

2.4. Surfactant-oil-water (SOW) phase behavior

The phase behavior of surfactant, oil, and water mixtures wasdetermined using a dilution experiment: increasing amounts ofwater were added to a surfactanteoil mixture. Initially, MCT wasmixed with surfactant (Tween 80, Tween 85, or 1:1 Tween 80/Tween 85) at a mass ratio of 1 part oil to 1 part surfactant.A series ofsamples with different water contents were then prepared byadding increasing amounts of double distilled water drop-wise intothe surfactanteoil mixtures (25�1 �C). The resulting SOW systemswere then mixed at a low speed using a magnetic stirrer to ensurethe different components were thoroughly mixed. After equilibra-tion, the appearance of the systems was observed visually and

using optical microscopy (Nikon Eclipse 80i, Nikon Instrument Inc.,Melville, NY).

3. Results and discussion

3.1. Comparison of emulsions formed by high- and low-energymethods

Initially, we examined the influence of homogenization methodand system composition on the formation of oil-in-water emul-sions. Our aimwas to establish the influence of these parameters onthe size of the droplets produced using the microfluidization andspontaneous emulsification methods, as well as to establish theminimum amount of surfactant needed to produce emulsionscontaining ultrafine droplets (r < 100 nm).

3.1.1. MicrofluidizaitonFor the microfluidization method, there was a steep decrease in

mean particle radius of the 20 wt% MCT oil-in-water emulsionswhen the surfactant concentration was increased from 0 to 2 wt%(SOW from 0 to 0.2), after which the particle size reached a rela-tively constant level (Fig. 2). The minimum particle radius attainedat high surfactant levels was 55 � 2 nm for all three systems(averaged over the data from 5 to 20 wt% surfactant). The depen-dence of particle size on surfactant concentration can be conve-niently divided into two distinct regimes for high pressurehomogenization (Tcholakova, Denkov, & Danner, 2004; Tcholakova,Denkov, Sidzhakova, Ivanov, & Campbell, 2003): the “surfactant-limited zone” and the “surfactant-rich zone”. At low surfactantconcentrations, there was insufficient surfactant present to coverthe new surfaces created within the homogenizer when largerdroplets were broken down to smaller ones (Fig. 1). Consequently,two or more small oil droplets that collided with each other wouldtend to coalesce because there was insufficient surfactant presentat their surfaces to prevent them from coming into close contact.Coalescence tends to continue until all of the droplets are coveredby sufficient surfactant to prevent them from merging together.Hence, the droplet size in the surfactant-limited zone is primarily

Fig. 3. Influence of surfactant concentration on the mean particle radius (r32) of 20 wt% MCT oil-in-water emulsions formed by a low-energy method (spontaneous emul-sification). The preparation procedure is described in the manuscript.

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406 401

determined by the type and amount of surfactant present. Theminimummean droplet radius (r32) that is obtained in this region isgiven by the following equation (McClements, 2005):

rmin ¼ 3$G$fcS

¼ 3$G$fc0Sð1� fÞ (1)

Here, G is the surface load of the surfactant (in kg m�2), f is thedisperse phase volume fraction, cS is the total surfactant concen-tration in the emulsion (in kg m�3) and c0S is the total surfactantconcentration in the continuous phase (in kg m�3). This equationassumes that all of the surfactant adsorbs to the droplet surfaces,and that the droplets do not grow after formation e.g., due to Ost-wald ripening or coalescence. In our study, the emulsions contained20 wt% oil droplets (f w 0.20) and 0.1 to 20 wt% surfactant in theaqueous phase (c0S w1e200 kg m�3). If we assume that all of thesurfactant adsorbed to the oilewater interface, then we can usethe measured values of rmin (55 nm) to predict the surface load ofthe surfactants. The value of c0S was taken to be the minimumsurfactant concentration required to reach the smallest dropletradius (i.e., the value at the end of the surfactant-limited zone),which was approximately 2 wt% (20 kg m�3). Consequently, thevalue of the surface load calculated for the surfactants usingEquation (1) was: G ¼ 1.5 mg m�2 (1.5 � 10�6 kg m�3), which isclose to published values for this parameter (McClements, 2005;Tcholakova, Denkov, & Lips, 2008). In practice, the surfactantmolecules will be distributed through many different environ-ments (e.g., oilewater interface; micelles and monomers in theaqueous phase: and, reverse micelles and monomers in the oilphase), and so not all of the surfactant will go to the oilewaterinterface. Nevertheless, this amount would be expected to berelatively small since the critical micelle concentrations of theTweens are typically very low<0.002 wt% (Hait & Moulik, 2001). Inaddition, the droplet surfaces may not need to be completelycovered by surfactant in order for protection against coalescence tooccur. The above value should therefore only be seen as a roughestimate of the actual surface load.

At relatively high surfactant concentrations there is sufficientsurfactant present to rapidly cover all of the small oil dropletscreated within the homogenizer, and so the droplet size is limitedby the maximum disruptive forces generated by the homogenizerrather than by surfactant properties. In this regime, adding moresurfactant therefore has little impact on the size of the dropletsproduced during homogenization. Previous studies have shownthat the particle size can be decreased in this regime by increasingthe homogenization pressure (P) to facilitate further dropletdisruption (Qian and McClements, 2011; Walstra, 1993; Woosteret al., 2008). Typically, there is a linear decrease in log(d) withincreasing log(P) for oil-in-water emulsions prepared using smallmolecule surfactants in a microfluidizer, with the slope dependingon homogenizer design.

In general, the two surfactants and their mixture showed similartrends in their particle size versus surfactant concentrationdependences, but there were some differences (Fig. 2). In thesurfactant-limited zone, the decrease in mean particle size withincreasing surfactant concentration was steeper for the Tween 80than for the Tween 85, and the mixture was somewhere inbetween. Thus, for the same surfactant concentration smallerdroplets were produced for Tween 80 than for Tween 85 in thesurfactant-limited zone. There are a number of possible explana-tions to account for this difference, e.g., there were differences inthe adsorption rate of the surfactants to the oilewater interface, orin the ability of the surfactants to stabilize the droplets againstcoalescence within the homogenizer. Tween 85 is more lipophilicthan Tween 80 and therefore therewill have been a greater fractionof this surfactant initially present within the oil phase. The oil phase

is more viscous than the aqueous phase and therefore Tween 85 inthe oil phase may have moved to the droplet interface more slowlythan Tween 80 in the water phase.

3.1.2. Spontaneous emulsificationThe spontaneous emulsification (SE) method was also used to

prepare 20 wt% MCT oil-in-water emulsions using the same typeand concentrations of surfactant (Fig. 3), so as to directly comparethe results with those obtained using the microfluidization method(Fig. 2). The influence of surfactant concentration on the fullparticle size distribution of the samples prepared using the SEmethod is also reported (Fig. 4). The SE method was capable offorming emulsions with small mean droplet radius (r < 100 nm),but the amount of surfactant required to form small droplets witha monomodal distribution was considerably higher than themicrofluidization method (Figs. 3 and 4). For example, a surfactant-to-oil ratio (SOR) of <0.1 was typically required to form smalldroplets (r < 100 nm) using the microfluidization method, buta SOR> 1 was required to form small droplets using the SEmethod.Thus, approximately ten-fold more surfactant was needed toproduce similarly-sized droplets using the low-energy approachthan using the high-energy approach. This clearly highlights one ofthe major disadvantages of the spontaneous emulsificationmethodfor forming food-grade ultrafine emulsions. For the Tween 80 andthe Tween 80/Tween 85 system, the particle size progressivelydecreased with increasing surfactant concentration with thesmallest mean droplet radius being obtained at 20% surfactant:r ¼ 77 nm for Tween 80 and r ¼ 55 nm for the Tween 80/Tween 85mixture. For the Tween 85 system there was a decrease in meanparticle radius from 0 to 8% surfactant, followed by an increase athigher surfactant concentrations (Fig. 3). A minimum dropletradius of around 87 nm was observed at 8% surfactant for thissystem. These results show that at low and intermediate surfactantconcentrations (<8%) the smallest lipid droplets can be formedwith Tween 85, but at higher surfactant concentrations the smallestdroplets can be formed with the Tween 80/Tween 85 mixture. Thefull particle size distributions showed that the droplets formedwere monomodal at sufficiently high surfactant concentrations for

Fig. 4. The particle size distributions of MCT oil-in-water emulsions formed by using different concentrations of surfactants prepared by the spontaneous emulsification method: (a)Tween 80; (b) Tween 80/Tween 85; (c) Tween 85.

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406402

the Tween 80 (Fig. 4a) and mixed systems (Fig. 4b), and at inter-mediate concentrations for the Tween 85 system (Fig. 4c).

The physicochemical origin of the relationship between dropletsize and system composition is much less well understood for lowenergy methods (such as spontaneous emulsification) than forhigh-energy methods (such as microfluidization). A large numberof empirical studies have examined the relationship betweenparticle size, system composition, and preparation conditions usinglow-energy methods (Salager et al., 2005; Salager et al., 2004; Solet al., 2006; Solans et al., 2003). Fernandez and co-workersproposed that ultrafine oil droplets were formed when water wasadded to a surfactant/oil mixture due to phase inversion: the

system went from a W/O microemulsion to a bicontinuous systemto an O/W emulsion (Fernandez, Andre, Rieger, & Kuhnle, 2004).Recently, Mercuri and co-workers proposed the following mecha-nism for the spontaneous formation of ultrafine oil droplets basedon a study of the phase behavior and microstructure of ternarymixtures of oil (soybean oil), surfactant (Tween 80 and Span 80)and water (Mercuri et al., 2011). When an aliquot of surfactant/oilmixture is brought into contact with a water phase a boundary isinitially formed between the S/O and W phases. Water thenpenetrates into the S/O phase causing it to swell, leading to theformation of W/O microemulsions and then to liquid crystallinephases. Eventually, a fragment (z50 mm) of the liquid crystalline

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406 403

S/O phase breaks off from the S/O-W boundary and moves into thewater phase. These relatively large fragments may then furtherdissociate into the nano-sized lipid droplets eventually formed.An alternative hypothesis proposed by Anton and co-workers toaccount for spontaneous emulsification is the movement ofsurfactant molecules from the S/O phase into the water phase afterthe two phases come into contact, which leads to spontaneousformation of nano-sized droplets at the boundary due to a buddingmechanism (Fig. 1) (Anton and Vandamme, 2009). In reality, it islikely that water will penetrate into the S/O phase and thatsurfactant will move into the water phase, although the relativerate and influence of these two processes is currently unknown.Nevertheless, neither of these mechanisms can currently be used toquantitatively account for the influence of system composition onthe size of the droplets formed by the SE method.

The fact that a minimum in droplet size was observed for theTween 85 system at around 8 wt% is currently not well understood.Previous studies have also reported a minimum in the droplet sizefor surfactant-oil-water systems containing Tween 85 and MCT(Pouton, 1985, 1997). These authors reported that the mean particlesize obtained using the SE method had a minimum value ata surfactant-to-oil ratio of Tween 85-to-MCT of about 3-to-7 (0.43-to-1), which is very similar to the value of 8-to-20 (0.40-to-1) foundin our work. Pouton and co-workers hypothesized that highconcentrations of this surfactant led to the formation of viscous gelsat the S/O-W boundary that retarded self-emulsification by limitingmolecular diffusion processes (Pouton,1985,1997). The influence ofsystem composition on the physicochemical characteristics of theSOW systems is reported in Section 3.3. It is clear that further workis required to rationalize the formation of oil-in-water emulsionswith specific droplet sizes using the self-emulsification approach.

3.2. Stability of the emulsion formed by spontaneous emulsification

The long-term stability of an emulsion is one of the mostimportant factors determining its suitability for application withinthe food and beverage industries (Dickinson, 1992; McClements,2005). For this reason, the stability of the emulsions produced inthis study was assessed bymeasuring changes in their droplet sizesand appearance during storage at ambient temperature (w20 �C)for onemonth. Emulsions formed usingmicrofluidization remainedstable to droplet aggregation (there was no change in their particlesize distribution) and gravitational separation (there was no visibleevidence of creaming) throughout a 30 day period, which can beattributed to their relatively small particle sizes (data not shown).

For the emulsions formed using the spontaneous emulsificationmethod, light scattering measurements indicated that the emul-sions stabilized by Tween 80 and by Tween 80/Tween 85 werestable to particle growth during one month storage, with nosignificant change in mean particle radius (Table 1) or particle sizedistribution (Fig. 5a). On the other hand, there was an appreciableincrease in the mean particle radius of the emulsions stabilized by

Table 1Surfactant properties and influence of surfactant type on mean particle radius ofemulsions formed by spontaneous emulsification after 0 and 30 days storage. Thesurfactant concentration is the valuewhere theminimumdroplet sizewas producedby the SE method. Data for the surfactants were taken from the literature (Hait andMoulik, 2001).

Surfactanttype

MW(g/mol)

HLB CMC(mg/100 g)

Surfactantconcentration

RadiusDay 0

RadiusDay 30

Tween 85 1310 15 1.31 8% 87 nm 183 nmTween 80 1839 11 0.033 20% 78 nm 77 nmTween 80:

Tween 851575 13 e 20% 55 nm 55 nm

Tween 85 during storage (Table 1), which can be attributed to anincrease in the fraction of larger droplets present in the particle sizedistribution (Fig. 5b). The observed differences in emulsion stabilitycan be attributed to differences in the influence of the molecularcharacteristics of the surfactants on droplet coalescence. It isknown that the rate of droplet coalescence tends to increase as thehydrophileelipophile balance (HLB) of non-ionic surfactantsdecreases from high to intermediate values (Tolosa, Forgiarini,Moreno, & Salager, 2006). Surfactants with high HLB numbers(>15) tend to be water soluble, form micelles, and stabilize oil-in-water emulsions, whereas surfactants with intermediate HLBnumbers (5e9) tend to be soluble in both the oil and water phases,form lamellar structures, and are not particularly good at stabilizingeither O/W or W/O emulsions (Israelachvili, 1992). The physico-chemical origin of this effect is that the interfacial tension at theoilewater interface tends to become very low as the HLB numbertends toward 7 (the surfactant curvature tends toward unity),which means the interface becomes highly mobile and susceptibleto coalescence through “hole” formation (Salager et al., 2003).

Visual observation of the 20 wt% MCT oil-in-water emulsionsafter 30 days storage indicated that homogenization method,

Fig. 5. Influence of storage at ambient temperature on the particle size distribution of20 wt% MCT oil-in-water emulsions formed by using different surfactants: (a) Tween80; (b) Tween 85. Like the Tween 80 system, the mixed surfactant system (Tween 80/Tween 85) was also stable to droplet growth during storage (data not shown).

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406404

surfactant type, and surfactant concentration also had a majorimpact on their stability to gravitational separation. As mentionedearlier, all the emulsions prepared using microfluidizationremained stable to gravitational separation during storage, whichmay be attributed to their relatively small droplet sizes (data notshown). A number of the samples prepared by the spontaneousemulsification method exhibited evidence of creaming instabilityafter storage, with an optically opaque cream layer being observedat the top of the samples and a slightly turbid serum layer at thebottom (Fig. 6). For the Tween 80 samples, the height of the creamlayer increased as the surfactant concentration increased, sug-gesting the samples were more stable to creaming at highersurfactant concentrations. For the Tween 85 samples, the emul-sions were fairly stable to creaming at low and intermediatesurfactant concentrations (4e12%), but showed extensive creamingat higher surfactant concentrations (16 and 20%). For the mixedsurfactant samples, the emulsions were unstable at the lowestsurfactant concentration (4%), were stable at intermediateconcentrations (8 and 12%), and showed some evidence ofcreaming again at the higher values (16 and 20%).

A number of factors may contribute to the creaming stability ofthe emulsions. First, the creaming velocity of a lipid dropletincreases as its radius increases, so that emulsions containing largerdroplets will cream more rapidly than those containing smaller

Fig. 6. Photographs of the influence of surfactant type and concentration on thestability of 20 wt% MCT oil-in-water emulsions produced by the spontaneous emul-sification method to gravitational separation after 30 days storage.

ones (McClements, 2005). Consequently, the emulsions preparedwith low surfactant concentrations are likely to be unstable tocreaming because they contain relatively large droplets. Theemulsions prepared with relatively low levels of Tween 85 hadsmaller droplet sizes than those prepared with Tween 80 or themixed system (Fig. 3), which may account for their better creamingstability at low surfactant concentrations (Fig. 6). On the otherhand, the relatively poor creaming stability of the Tween 85samples at high surfactant concentrations (Fig. 6) may be due totheir relatively large particle size (Fig. 3). Second, creaming may beinduced at intermediate surfactant levels due to the presence ofnon-adsorbed surfactant micelles that generate a depletionattraction between the droplets (McClements, 1994). The strengthof the depletion attraction increases with increasing surfactantconcentration, and above a critical surfactant level (which increaseswith decreasing droplet size) flocculation may be induced, leadingto rapid creaming. This phenomenon may account for the fact thatsome creaming was observed in the Tween 85 and mixed surfac-tant systems at high surfactant concentrations, even though theyhad relatively small droplet sizes.

3.3. Phase behavior of SOW systems

The phase behavior of the system at the oil/surfactant-waterboundary plays a key role in the formation of lipid droplets bythe spontaneous emulsification process (Salager et al., 2004). Wetherefore used dilution experiments and optical microscopymeasurements to provide some information about the phasebehavior of the SOW systems used in this study.

Fig. 7. Photographs of the influence of surfactant type on the appearance and prop-erties of surfactant-oil-water (SOW) mixtures containing different SO-W ratios, buta fixed SO ratio (1:1). Annotation: the numbers x/y represent the mass ratio of SO-to-W, e.g., 9/1 means there was 9 g of surfactant-oil mixture and 1 g of water.

Fig. 8. Micrographs of the influence of surfactant type on the appearance of surfactant-oil-water (SOW) mixtures containing different SO-W ratios, but a fixed SO ratio (1:1).Annotation: the numbers “x/y” represent the mass ratio of surfactanteoil (SO) to water (W), e.g., 9/1 means there was 9 g of surfactanteoil mixture and 1 g of water. The small whitebar in the lower right hand corner of each image represents 10 mm.

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406 405

A series of mixtures with different mass fractions of surfactant(S), oil (O), and water (W), but with a fixed surfactant-to-oil ratio(S:O¼ 1:1) was prepared, and then their overall appearance (Fig. 7)and microstructure (Fig. 8) was measured. Visual observationssuggested that all of the surfactants exhibited relatively similarbehavior when the amount of water present in the system wasincreased: at high SO-to-W ratios (low water contents) a viscoustransparent oily solution was formed; at intermediate SO-to-Wratios a gel-like translucent material was formed; at high SO-to-W ratios (high water contents) a low viscosity milky liquid wasformed (Fig. 7). Examination of the SOW systems using opticalmicroscopy indicated that surfactant type did have an appreciableinfluence on their microstructures (Fig. 8). The Tween 80 samplewas relatively homogeneous at low and high SO-to-W ratios, sug-gesting that large particles were not present in the oil or watercontinuous systems. At high SO-to-W ratios, this corresponds to thetransparent oily liquid observed at low water contents. At low SO-to-W ratios, this corresponds to the milky liquid formed at highwater contents. A highly heterogeneous microstructure wasobserved under the microscope at intermediate SO-to-W ratios i.e.7:3 (Fig. 8), which corresponded to the gel-like structure observedvisually (Fig. 7). The Tween 85 showed quite different behavior,with large heterogeneous structures being formed at all surfactant/oil to water ratios studied (Fig. 8). The precise nature of thesestructures is currently unknown, although it is likely that they aresome kind of liquid crystalline phase. Large heterogeneous struc-tures were only observed in the Tween 80/85 system at a SO-to-Wratio of 8:2 (data not shown), otherwise there was no large struc-tures formed (Fig. 8). Observation of the same systems using cross-polarized optical microscopy indicated that they had some order,i.e., the structures rotated polarized light and gave a whiteappearance against a black background (data not shown). This

birefringent behavior suggests that the surfactants formed liquidcrystalline structures at intermediate SO-to-W ratios.

4. Conclusions

This study has highlighted some of the advantages and disad-vantages of low- and high-energy approaches for creating oil-in-water emulsions containing ultrafine oil droplets. The high-energy method (microfluidization) was able to produce emulsionswith small droplets (r < 100 nm) using low surfactant-to-oil ratios(<1:10), but it required relatively expensive specialized equipmentand has relatively high operating costs. On the other hand, the low-energy method (spontaneous emulsification) is inexpensive andcan be implemented using simple equipment (stirring), but itrequires high surfactant-to-oil ratios (>5:10) to produce smalldroplets. The usage of relatively high surfactant levels may causeproblems in some food and beverage products due to high ingre-dient costs, off flavors, and safety concerns. Further work isrequired to establish whether low-energy methods can be devel-oped that produce small droplet sizes without the need for highsurfactant levels.

References

Acosta, E. (2009). Bioavailability of nanoparticles in nutrient and nutraceuticaldelivery. Current Opinion in Colloid & Interface Science, 14(1), 3e15.

Anton, N., Benoit, J., & Saulnier, P. (2008). Design and production of nanoparticlesformulated from nano-emulsion templatesea review. Journal of ControlledRelease, 128(3), 185e199.

Anton, N., & Vandamme, T. F. (2009). The universality of low-energy nano-emul-sification. International Journal of Pharmaceutics, 377(1e2), 142e147.

Dickinson, E. (1992). An introduction to food colloids. Oxford: Oxford SciencePublishers.

Y. Yang et al. / Food Hydrocolloids 29 (2012) 398e406406

Einhorn-Stoll, U., Weiss, M., & Kunzek, H. (2002). Influence of the emulsioncomponents and preparation method on the laboratory-scale preparation ofo/w emulsions containing different types of dispersed phases and/or emulsi-fiers. Nahrung/Food, 46(4), 294e301.

Fernandez, P., Andre, V., Rieger, J., & Kuhnle, A. (2004). Nano-emulsion formation byemulsion phase inversion. Colloids and Surfaces A-Physicochemical and Engi-neering Aspects, 251(1e3), 53e58.

Hait, S. K., & Moulik, S. P. (2001). Determination of critical micelle concentration(CMC) of nonionic surfactants by donor-acceptor interaction with iodine andcorrelation of CMC with hydrophileelipophile balance and other parameters ofthe surfactants. Journal of Surfactants and Detergents, 4(3), 303e309.

Henry, J. V. L., Fryer, P. J., Frith, W. J., & Norton, I. T. (2009). Emulsification mechanismand storage instabilities of hydrocarbon-in-water sub-micron emulsions sta-bilised with Tweens (20 and 80), Brij 96v and sucrose monoesters. Journal ofColloid and Interface Science, 338(1), 201e206.

Henry, J. V. L., Fryer, P. J., Frith, W. J., & Norton, I. T. (2010). The influence of phos-pholipids and food proteins on the size and stability of model sub-micronemulsions. Food Hydrocolloids, 24(1), 66e71.

Horn, D., & Rieger, J. (2001). Organic nanoparticles in the aqueous phase-theory,experiment, and use.Angewandte Chemie-International Edition, 40(23), 4331e4361.

Israelachvili, J. (1992). Intermolecular and surface forces (2nd ed.). London, UK:Academic Press.

Jafari, S., He, Y., & Bhandari, B. (2007). Optimization of nano-emulsions productionby microfluidization. European Food Research and Technology, 225(5), 733e741.

Kentish, S., Wooster, T., Ashokkumar, M., Balachandran, S., Mawson, R., & Simons, L.(2008). The use of ultrasonics for nanoemulsion preparation. Innovative FoodScience & Emerging Technologies, 9(2), 170e175.

McClements, D. J. (1994). Ultrasonic determination of depletion flocculation in oil-in-water emulsions containing a nonionic surfactant. Colloids and SurfacesA-Physicochemical and Engineering Aspects, 90(1), 25e35.

McClements, D. J. (2002). Theoretical prediction of emulsion color. Advances inColloid and Interface Science, 97(1e3), 63e89.

McClements, D. J. (2005). Food emulsions: Principles, practice, and techniques. BocaRaton: CRC Press.

McClements, D. J. (2010). Emulsion design to Improve the delivery of functionallipophilic components. Annual Review of Food Science and Technology, 1(1),241e269.

McClements, D. J. (2011). Edible nanoemulsions: fabrication, properties, and func-tional performance. Soft Matter.

McClements, D. J., Decker, E. A., & Weiss, J. (2007). Emulsion-based delivery systemsfor lipophilioc bioactive components. Journal of Food Science, 72(8), R109eR124.

McClements, D. J., & Rao, J. (2011). Food-grade nanoemulsions: formulation, fabri-cation, properties, performance, biological fate, and potential toxicity. CriticalReviews in Food Science and Nutrition, 51(4), 285e330.

Mason, T. G., Wilking, J. N., Meleson, K., Chang, C. B., & Graves, S. M. (2006).Nanoemulsions: formation, structure, and physical properties. Journal ofPhysics-Condensed Matter, 18(41), R635eR666.

Mercuri, A., Passalacqua, A., Wickham, M. S. J., Faulks, R. M., Craig, D. Q. M., &Barker, S. A. (2011). The effect of composition and gastric conditions on the self-emulsification process of ibuprofen-loaded self-emulsifying drug deliverysystems: a microscopic and dynamic gastric model study. PharmaceuticalResearch, 28(7), 1540e1551.

Miller, C. A. (1988). Spontaneous emulsification produced by diffusionea review.Colloids and Surfaces, 29(1), 89e102.

Nielsen, F. S., Gibault, E., Ljusberg-Wahren, H., Arleth, L., Pedersen, A. S., &Mullertz, A. (2007). Characterization of prototype self-nanoemuisifyingformulations of lipophilic compounds. Journal of Pharmaceutical Sciences,96(4), 876e892.

Nielsen, F. S., Petersen, K. B., & Mullertz, A. (2008). Bioavailability of probucol fromlipid and surfactant based formulations in minipigs: influence of droplet sizeand dietary state. European Journal of Pharmaceutics and Biopharmaceutics,69(2), 553e562.

Perrier-Cornet, J., Marie, P., & Gervais, P. (2005). Comparison of emulsificationefficiency of protein-stabilized oil-in-water emulsions using jet, high pressureand colloid mill homogenization. Journal of Food Engineering, 66(2), 211e217.

Pouton, C. W. (1985). Self-emulsifying drug delivery systems: assessment of theefficiency of emulsification. International Journal of Pharmaceutics, 27, 335e348.

Pouton, C. W. (1997). Formulation of self-emulsifying drug delivery systems.Advanced Drug Delivery Reviews, 25, 47e58.

Pouton, C. W., & Porter, C. J. H. (2006). Formulation of lipid-based delivery systemsfor oral administration: materials, methods and strategies. In. Annual Meeting ofthe American-Association-of-Pharmaceutical-Scientists, San Antonio, TX.

Qian, C., & McClements, D. J. (2011). Formation of nanoemulsions stabilized bymodel food-grade emulsifiers using high-pressure homogenization: factorsaffecting particle size. Food Hydrocolloids, 25(5), 1000e1008.

Rao, J. J., & McClements, D. J. (2010). Stabilization of phase inversion temperaturenanoemulsions by surfactant Displacement. Journal of Agricultural and FoodChemistry, 58(11), 7059e7066.

Sadurní, N., Solans, C., Azemar, N., & García-Celma, M. (2005). Studies on theformation of O/W nano-emulsions, by low-energy emulsification methods,suitable for pharmaceutical applications. European Journal of PharmaceuticalSciences, 26(5), 438e445.

Sagalowicz, L., & Leser, M. E. (2010). Delivery systems for liquid food products.Current Opinion in Colloid & Interface Science, 15(1e2), 61e72.

Salager, J.-L., Forgiarini, A., Marquez, L., Pena, A., Pizzino, A., Rodriguez, M. P., et al.(2004). Using emulsion inversion in industrial processes. Advances in Colloidand Interface Science, 108e109, 259e272.

Salager, J., Antón, R., Sabatini, D., Harwell, J., Acosta, E., & Tolosa, L. (2005).Enhancing solubilization in microemulsionsdstate of the art and currenttrends. Journal of Surfactants and Detergents, 8(1), 3e21.

Salager, J. L., Marquez, L., Mira, I., Pena, A., Tyrode, E., & Zambrano, N. B. (2003).Principles of emulsion formulation engineering. In Mittal, K. L., & Shah, D. O.(Eds.). (2003). Adsorption and aggregation of surfactants in solution, Vol. 109 (pp.501e523).

Sanguansri, P., & Augustin, M. A. (2006). Nanoscale materials development e a foodindustry perspective. Trends in Food Science & Technology, 17(10), 547e556.

Sol, I., Maestro, A., Pey, C., González, C., Solans, C., & Gutiérrez, J. (2006). Nano-emulsions preparation by low energy methods in an ionic surfactant system.Colloids and Surfaces A: Physicochemical and Engineering Aspects, 288(1e3),138e143.

Solans, C., Esquena, J., Forgiarini, A. M., Uson, N., Morales, D., Izquierdo, P., et al.(2003). Nano-emulsions: formation, properties, and applications. In Mittal, K. L.,& Shah, D. O. (Eds.). Adsorption and aggregation of surfactants in solution, Vol. 109(pp. 525e554).

Sonneville-Aubrun, O., Simonnet, J. T., & L’Alloret, F. (2004). Nanoemulsions: a newvehicle for skincare products. Advances in Colloid and Interface Science, 108,145e149.

Tadros, T., Izquierdo, R., Esquena, J., & Solans, C. (2004). Formation and stabilityof nano-emulsions. Advances in Colloid and Interface Science, 108e09,303e318.

Tcholakova, S., Denkov, N. D., & Danner, T. (2004). Role of surfactant type andconcentration for the mean drop size during emulsification in turbulent flow.Langmuir, 20(18), 7444e7458.

Tcholakova, S., Denkov, N. D., & Lips, A. (2008). Comparison of solid particles,globular proteins and surfactants as emulsifiers. Physical Chemistry ChemicalPhysics, 10(12), 1608e1627.

Tcholakova, S., Denkov, N. D., Sidzhakova, D., Ivanov, I. B., & Campbell, B. (2003).Interrelation between drop size and protein adsorption at various emulsifica-tion conditions. Langmuir, 19(14), 5640e5649.

Tolosa, L. I., Forgiarini, A., Moreno, P., & Salager, J. L. (2006). Combined effects offormulation and stirring on emulsion drop size in the vicinity of three-phasebehavior of surfactant-oil water systems. Industrial & Engineering ChemistryResearch, 45(11), 3810e3814.

Velikov, K. P., & Pelan, E. (2008). Colloidal delivery systems for micronutrients andnutraceuticals. Soft Matter, 4(10), 1964e1980.

Walstra, P. (1993). Principles of emulsion formation. Chemical Engineering Science,48, 333.

Wooster, T., Golding, M., & Sanguansri, P. (2008). Impact of oil type on nano-emulsion formation and Ostwald ripening stability. Langmuir, 24(22),12758e12765.