-o

Keywords:

Sorbitan monooleateLecithin

mulr-ine stmb

emulsiers (PGPR, Span 80 and lecithin) at two water:oil ratios was investigated. Emulsions with higher

pharmin-oilsions d

a three-dimensional network (McClements, 2005; Rousseau, 2000).However, uid W/O emulsions generally present low stability

because of the high mobility of water droplets, which can easilysediment, occulate or coalesce. A number of studies about emul-sions focuses on O/W systems, but there are fewworks about liquidW/O emulsions. A better understanding about the interactions

thanolamine (PE)id (PA) and otherlecithin is similar

to that of the soybean oil (Bergensthl & Fontell, 1983). Lecithin issoluble in mineral oils, but practicable insoluble in non-heatedvegetable oils. It is also insoluble but highly dispersible in water(Wendel, 2000). Moreover, a powerful hydrophobic emulsier usedin food industry is the polyglycerol polyricinoleate (PGPR), anoligomeric and non-ionic emulsier (Fig. 2a) produced by theesterication of castor oil fatty acid (Fig. 2b) with polyglycerol. Withlecithin, PGPR is widely used to reduce the viscosity in chocolateproducts (Weyland & Hartel, 2008). Other food-grade emulsier

* Corresponding author. Tel.: 55 19 3521 4047; fax: 55 19 3521 4027.

Contents lists available at

r

els

Food Hydrocolloids 34 (2014) 145e153E-mail address: rosiane@fea.unicamp.br (R.L. Cunha).based products are in the solid or semi-solid state, such as butterand margarine, and are stabilized by fat crystallization, whichprevents the sedimentation of water droplets by the formation of

dylcholine (PC) (Fig. 1a), followed by phosphadyle(Fig. 1b), phosphatidylinositol (PI), phosphatidic acminorities. The fatty acid composition of soybeanwater (O/W) emulsions, which can be stabilized by both stericand electrostatic repulsion. In the case of W/O emulsions, onlysteric forces are expected to stabilize the emulsion, because of thelow electrical conductivity of the continuous phase (Claesson,Blomberg, & Poptoshev, 2004). In fact, most of the W/O emulsion

hydrophileelipophile balance (HLB). Among the hydrophobic food-grade emulsiers, lecithin is the only natural one (Dickinson, 1993).It is an emulsier with amphoteric character contained in differentnatural sources, such as soybean and egg, and is composed bya mixture of phospholipids. Its main component is the phosphati-1. Introduction

There are many products in thefood industry that contain water-stability mechanism of those emul0268-005X/$ e see front matter 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.foodhyd.2012.11.016hindered by a stable elastic interface. The molecular structure of both oil and emulsier were importantto dene the emulsion stability. Better chemical afnity of the hydrophobic moieties of the emulsierand the oil led to more stable interface. Steric stability was obtained in more viscous systems, such asthose at higher water volume fraction content. However, the water incorporation capacity into theemulsion depended on the molecular structure of hydrophilic portion of the emulsier. Moreover, thewater e soybean oil systems with Span 80 or lecithin emulsiers did not form a macroemulsion, buta gelled structure. This study discussed the many factors that affect the emulsion formation and stability,which can contribute to the development of new water-in-oil emulsion based products with higherstability.

2012 Elsevier Ltd. All rights reserved.

aceutical, cosmetic and(W/O) emulsions. Theiffers from the oil-in-

between water, oil and emulsier at the interface and the factorsthat affect the emulsion stability would allow producing stableliquid W/O emulsions and, therefore, would encourage the devel-opment of new products and applications.

The emulsiers used to prepare W/O emulsions have lowEmulsion stabilityPolyglycerol polyricinoleateSpan 80 e presented an interface with low initial interfacial tension and practically constant complexviscoelastic modulus with time. Therefore, small droplets were formed and their coalescence wasWater-in-oil emulsion kinetic stability e water and soybean oil emulsion stabilized with PGPR and water and hexadecane withStability mechanisms of liquid water-in

F.Y. Ushikubo, R.L. Cunha*

Faculty of Food Engineering, University of Campinas, 13.083-862 Campinas, SP, Brazil

a r t i c l e i n f o

Article history:Received 2 June 2012Accepted 12 November 2012

a b s t r a c t

Although the stability of ewater emulsions or watestabilize. In this study, thdifferent systems. The co

Food Hyd

journal homepage: www.All rights reserved.il emulsions

sions is widely discussed in the literature, most of them dealt with oil-in--oil systems with solid and semi-solid structures, which are easier toability mechanism of liquid water-in-oil emulsions was investigated inination of two different oils (soybean oil and hexadecane) and three

SciVerse ScienceDirect

ocolloids

evier .com/locate/ foodhyd

OO

O

RO

O P O

N+(CH3)3

O-

OH

OO

O

RO

O P O

O-

OHNH3

+

atid

F.Y. Ushikubo, R.L. Cunha / Food Hydrocolloids 34 (2014) 145e153146with low HLB is sorbitan fatty acid esters, which are monomericand non-ionic emulsiers. Particularly, sorbitan monooleate (Span80) is obtained from a reaction between sorbitol and fatty acids(Fig. 3). PGPR and Span 80 are soluble in mineral and vegetable oils,but insoluble in water. All those emulsiers are commonly used inthe food industry, such as in bakery, dairy, confectionary and oilprocessing (Bueschelberger, 2006; Cottrell & van Peij, 2006), inthe cosmetic industry, such as skin care creams, lip balms,sunscreens, shampoos and soaps, and in pharmaceuticals (Tadros,2008; Wendel, 2000).

Furthermore, the bulk physicochemical characteristics are alsoimportant for the formation and stability of the emulsion. Forexample, the density is decisive in the stability to sedimentation,the viscosity has inuence in the droplets breakup and the mobilityof droplets, and the oil polarity can affect the interfacial tensionwith the aqueous phase and the partitioning of the components atthe interface. Besides, the chemical structure, such as fatty acidschain length, number of unsaturations, molecule congurationhave also effect in the stability of the emulsion. Many studies usemodel systems with highly puried oils and with known chemicalcomposition, but the behavior of those systems is different from thecomplex food-grade oils (McClements, 2005).

Thus, this study is aimed to evaluate the stability of W/Oemulsions formulated with different types of oil and emulsiers attwo water:oil ratios. In order to understand the stability mecha-nisms, the interface of the systems was studied regarding theinterfacial tension and the viscoelastic properties. Besides, thestructure of the emulsions was analyzed microscopically throughimages and particle size distribution, and macroscopically, by therheology of the emulsions.

2. Materials and methods

Oils with different characteristics were chosen to prepare theemulsions: n-hexadecane (C16H34) (Sigma Aldrich, USA), a linearchain hydrocarbon (molecular weight, MW 226.4 g mol1), withvery low polarity and viscosity (3 mPa s); and soybean oil (Soya,Bunge Alimentos S.A., Brazil), a triglyceride composed mainly by43e56% linoleic (C18:2), 15e33% oleic (C18:1), 7e11% palmitic (C16:0)and 5e11% a-linolenic (C18:3) acids (Poth, 2001), (average

(a)R

Fig. 1. Formula of (a) phosphatidylcholine and (b) phosphMW 278.2 g mol1) with higher polarity and viscosity (50 mPa s)than the former oil. Deionized water was used as aqueous phase.

The emulsiers were sorbitan monooleate (Span 80), HLB4.3, MW 428.6 g mol1, kindly provided by Croda do Brasil

(a)

RO OR

OR

n

Fig. 2. (a) Formula of PGPR, in which R is a hydrogen, ricinolei(Brazil); natural lipophilic lecithin (Solec SG), HLB 4, averageMWy 750 g mol1, with phospholipid composition of 16% (w/w)phosphadidylcholine (PC), 14% phosphatidylethanolamine (PE), 9%phosphatidylinositol (PI) and 5% phosphatidic acid (PA), accordingto the supplier (Solae, Brazil) and polyglycerol polyricinoleate(GRINDSTED PGPR), HLB 1.5e2.0, average MW y 3000 g mol1,donated by Danisco Brasil Ltda (Brazil).

2.1. Emulsion preparation

Emulsions were prepared at 30:70 and 60:40 (wwater/woil) ratios,aiming to evaluate the effect of the volume fraction of water inthe W/O emulsion stability. In order to observe the interactionbetween the emulsier and the oil, each emulsier (Span 80, leci-thin and PGPR) was used to stabilize water-in-oil (W/O) emulsionscomposed bywater as aqueous phase and n-hexadecane or soybeanoil as the oil phase.

For the emulsion preparation, the emulsier was previouslydissolved into the oil. Then, the aqueous phasewas added dropwiseto the oil phase while homogenizing in a rotor-stator system (UltraTurrax T18, IKA, Germany) at 14,000 rpm. When the aqueous phasewas completely incorporated into the emulsion (after 10 min at30:70 water:oil ratio and after 20 min at 60:40), the rotationalspeed was decreased to 11,000 rpm and the system was homoge-nized for 4 min. The emulsions were prepared at room temperature(25 1 C).

2.2. Interfacial tension and dilational rheology

The interfacial tension was measured at ambient temperature(25 1 C) using the pendant drop method using a Tracker-Stensiometer (Teclis Sarl, France), in which the deionized waterwas injectedwith a syringe at a volume of 5 ml and 1.5 ml for soybeanoil and hexadecane systemswith emulsiers, respectively, and 25 mlfor oil pure systems. For the calculation of the viscoelastic proper-ties, an oscillation of 10% of the volume was set at a frequency of0.2 Hz.

2.3. Evaluation of kinetic stability

(b)R

ylethanolamine, in which RCOOe are fatty acid residues.The phase separation of the emulsions was observed in a 25 mLcylindrical tube (internal diameter 17 mm), sealed with a plasticcap and stored for 14 days at 25 C. The oil phase separation wasexpressed in sedimentation index (SI), which is calculated as the

(b)

OH

O

OH

c acid or polyricinoleic acid. (b) Formula of ricinoleic acid.

medium. The measurements were done at laser obscuration

Pni$d4i

s s0 k _gn; (4)

in which s is the shear stress (Pa), s0 is the yield stress (Pa), _g is theshear rate (s1), k is the consistency index (Pa sn) and n is the owindex.

2.7. Statistical analysis

The difference in the results of particle size distribution andrheological parameters was statistically analyzed through theTukey test, using the software Statistica 7.0 (Statsoft Inc., USA).

3. Results

3.1. Interfacial tension

Hydrocolloids 34 (2014) 145e153 147d43 Pni$d3i

; (2)

in which n is the number of d diameter droplets.The polydispersion of the size distribution as evaluated through

the span value, shown in Eq. (3):

span d90 d10; (3)between 5 and 15% and at 1750 rpm agitation speed. The particlesize distribution was presented in volume fraction, considering theparameters of water droplets (refractive index 1.33 andabsorption 0.01).

The particle size was represented as volume-weighted meandiameter (d43), dened by Eq. (2):

relation between the upper oil phase height (H) and the initialheight (H0), as described by Eq. (1):

SI % H=H0*100 (1)In the case of emulsions in which the aqueous phase separated

from the emulsion, the percentage of aqueous phase volume inrelation to the total volume was calculated.

2.4. Optical microscopy

The emulsion microstructure was observed through an opticalmicroscope (Axio Scope.A1, Carl Zeiss, Germany) at 100 magni-cation for qualitative evaluation. The images were captured withthe software AxioVision Rel. 4.8 (Carl Zeiss, Germany).

2.5. Particle size distribution

The particle size distribution was determined by the laserdiffraction method using a Mastersizer 2000 (Malvern InstrumentsLtd., UK). The emulsions were dispersed in their respectivecontinuous phase oil: soybean oil and hexadecane (refractiveindexes 1.467 and 1.435, respectively). In order to avoid the pres-ence of bubbles, ultrasound was applied for 5 min in the dispersion

Fig. 3. Formula of span 80.F.Y. Ushikubo, R.L. Cunha / Foodd50

inwhich d10 d50 and d90 are the particle diameter belowwhich 10%,50% and 90% of the sample lies, respectively.

2.6. Rheological essays

Flow curves were obtained by a stress-controlled rheometer(Physica MCR301, Anton Paar, UK) at 25 C. The shear rate varied inthe range between 0 and 300 s1 and the ow curves wereobtained in sequential three steps: up-down-up. The second upcurve was tted to the HerscheleBulkley model:The interfacial tension measured at the waterepure soybean oilinterface was about 22 mN/m, while at waterehexadecane inter-face was 40 mN/m. When the emulsiers were added, all of themshowed a considerable reduction of the interfacial tension (Fig. 4).At wateresoybean oil interface, the initial interfacial tension wasaround 6.7 mN/m with PGPR and 2.9 mN/m with Span 80. In thelatter system, an opaque lm formed at the interface during thetension measurement. The rigid interface of wateresoybean oilstabilized with Span 80 could have contributed to lower theinterfacial tension. The interfacial tension of both systemsdecreased with time, reaching values of 2.7 mN/m with PGPRand 2.3 mN/m with Span 80 in the stability. The initial interfacialtension at waterehexadecane interface was approximately5.9 mN/m with PGPR and 6.1 mN/m with Span 80 and after 1 hdecreased to 3.2 and 3.6, respectively. Lecithin showed low inter-facial tension at wateresoybean oil interface (1.2 mN/m) atstability, although the initial tensionwas higher (9.2 mN/m) than inother emulsiers systems. At waterehexadecane interface withlecithin, the initial interfacial tensionwas even higher (24.1 mN/m),thus the mobility of lecithin was much lower. The interfacialtension of that system at the stability was 2.8 mN/m. Comparing allthe systems, the interfacial tension of waterehexadecane in thestability condition was higher than that of wateresoybean inter-face, due to the less polar character of hexadecane.

3.2. Interfacial dilational rheology

The complex viscoelastic modulus of the wateresoybean oilwith PGPR showed a slight decrease along time, keeping a value of

1

2

3

4

5

6

7

0 500 1000 1500 2000 2500 3000 3500 4000

In

terfacial Ten

sio

n, m

N/m

Time, s

05

10152025

0 500 1000 1500 Int. Ten

sio

n, m

N/m

Time, s

Fig. 4. Interfacial tension at soybean oilewater (full symbols) and hexadecaneewater(open symbols) interfaces with (A) PGPR, (-) Span 80 and (C) lecithin. The detail

shows the curve with larger scale related to the hexadecaneewater interface withlecithin.

accentuated at 30:70 than at 60:40 W:O ratio. In turn, lecithinemulsions produced with soybean oil showed slower phase sepa-ration rate than emulsions with Span 80, mainly at 30:70W:O ratio.However, differently from Span 80 emulsions, the phase separationin lecithin systems did not stabilize with time. In the last observa-tion day, the 30:70 sample had higher SI than the equivalent Span80 system (Table 1). Moreover, at 60:40W:O ratio, the separation ofthe aqueous phase indicated that lecithin could not stabilize thewater droplets, oppositely to PGPR and Span 80, which showedhigher stability in emulsions at higher water content.

studied in microscopic and rheological approaches.

F.Y. Ushikubo, R.L. Cunha / Food Hydrocolloids 34 (2014) 145e153148approximately 5 mN/m (Fig. 5). This value was the highest amongthe evaluated systems. At waterehexadecane interface with PGPR,the complex modulus increased with time from about 1.5 mN/m to4mN/m. On the other hand, Span 80 at wateresoybean oil interfaceled to a steep decrease in complex viscoelastic modulus with time,reaching values near zero. At waterehexadecane interface, a lowviscoelastic modulus (about 2 mN/m) was kept constant alongtime. The modulus at the wateresoybean oil interface with lecithindecreased from a high value (about 35 mN/m) to decrease up tovalues near 2mN/m. The results of thewaterehexadecane interfacewith lecithin oscillated between 20 and 100 mN/m, although thoseresults are not so reliable, since the surface area of the dropextensively changed with time (data not shown).

3.3. Kinetic stability

The stability of the emulsions was observed for 14 days. Mostemulsions destabilized only through sedimentation mechanism.However, lecithin formulations with high water content showedaqueous phase separation, resulting in a three-phase system. Theappearance of the emulsion and the sedimentation index (SI) andaqueous phase separation at the equilibrium are presentedin Table 1. The oil phase separation along time is shown in Fig. 6aand b, which presents the data of emulsions preparedwith soybeanoil and hexadecane, respectively.

After the 14-day storage, each oilesurfactant combinationresulted in products with distinct appearance. Those prepared withPGPR were a homogeneous liquid. Similar aspect was observed inall waterehexadecane systems. However, the wateresoybean oilemulsions with Span 80 or lecithin showed a gel-like structure. Inthe Span 80 system at 60:40 W:O ratio the structure was a rmer

Fig. 5. Complex dilational viscoelastic modulus at soybean oilewater (full symbols)and hexadecaneewater (open symbols) interfaces with (A) PGPR, (-) Span 80 and(C) lecithin. Complex elastic modulus of the hexadecaneewaterelecithin system isnot shown because it is out of the axis range.gel (Table 1).The emulsions produced with PGPR and soybean oil were the

most stable, resulting in low SI at both W:O ratios through thewhole observation time (Table 1, Fig. 6a). The systems with Span 80and soybean oil presented a faster phase separation followed bya tendency to stabilize with time. The phase separation was more

Table 1Appearance of W/O emulsions and Sedimentation Index (SI) and the aqueous phase sep

Type of oil Parameter 30:70

PGPR Span 80

Soybean oil Appearance Liquid GelSI 2.8% 42.0%AS e e

Hexadecane Appearance Liquid LiquidSI 68.4% 62.8%AS e e3.4. Optical microscopy imaging and particle size distribution

The microscopic images of the emulsions with soybean oil justafter the homogenization are shown in Fig. 7. It is noted thatemulsions prepared with PGPR are composed of spherical smalldroplets. On the other hand, the systems composed of Span 80presents a different structure than an emulsion, with some non-spherical particles. Similar appearance was observed at both W:Oratios. In the case of wateresoybean oil systems with lecithin, somebig gel-like structures were formed in the continuous phase. Thosestructures agglomerated during the observation of the microscopyimage, showing the high instability of the system.

Differently from the soybean oil emulsions, the microscopicimages of emulsions produced with hexadecane showed theformation of spherical droplets in all samples (Fig. 8). In general, thewaterehexadecaneePGPR emulsions showed larger droplets thanemulsions with soybean oil and the same emulsier. However,waterehexadecaneeSpan 80 emulsions showed smaller dropletsthan PGPR emulsions. At 30:70 W:O ratio, the droplets seem to beocculated. The droplets of the emulsion with lecithin were very

aration (AS) after 14-day storage.

60:40

Lecithin PGPR Span 80 Lecithin

Gel Liquid Gel Gel49.6% 2.4% 16.0% 29.6%e e e 38.4%Liquid Liquid Liquid Liquid72.4% 24.4% 9.2% 46.0%In general, the phase separation of emulsions prepared withhexadecane was much faster than in the case of the emulsionsprepared with soybean oil, mainly at 30:70 W:O ratio, in which thesedimentation index was higher than 40% in less than one day(Fig. 6b). The emulsions with hexadecane and lecithin were themost unstable, since practically instantaneous phase separationwas observed. Besides, similarly to the soybean oil emulsion withthe same emulsier, there was aqueous phase separation in theemulsion at 60:40 W:O ratio. The PGPR emulsion with hexadecaneat 60:40 W:O ratio showed better stability than the respectivesystem at 30:70, but much worse than the soybean oil systems.

On the other hand, the emulsion prepared with water andhexadecane stabilized by Span 80 at 60:40W:O ratio showed betterkinetic stability than the respective system with soybean oil. Aftertwo days of storage, only 0.8% of the hexadecane separated fromthe emulsion, while in thewateresoybean oil system, 8.0% of the oilhad already come apart from the emulsion.

In order to understand the stabilization mechanism and thecharacteristics of each watereoil system formed using different oilsand emulsiers, the structure and viscosity of the emulsions wase e e 18.0%

large. At 60:40 W:O ratio, the droplets are coalescing, forminga bicontinous phase.

The particle size distribution of emulsions prepared withsoybean oil and hexadecane is shown in Fig. 9a and b, respectively.The volume-weighted mean diameter (d43) and the span value arepresented in Table 2.

From those data, it is clearly seen that emulsions produced withPGPR and soybean oil presented the smallest particles among thesystems, showing a bimodal distribution. In PGPR samples at bothW:O ratios, the rst peak mode was around 0.1e0.2 mm, whichcould represent the excess of emulsier that is free or as smallaggregates in the continuous phase. This interpretation is corrob-

at larger particle sizes (Fig. 9a). Although the frequency of peaksof small size is low, they cannot be neglected, since the distributionwas plotted in volume frequency. From the particle size distributionand the values of d43 and span, it was observed that the particlesof samples at 30:70 were smaller and less polydisperse than at60:40 W:O ratio (Table 2). Based on the optical microscopyimages, in which no spherical droplets were observed, in bothsamples with Span 80 and lecithin, the particle size distributionprobably represents the distribution of aggregated water dropletsand/or small particles of the gel.

Regarding the emulsions with hexadecane, the PGPR and Span80 emulsions showed a monomodal distribution with low poly-

Fig. 6. Sedimentation index (SI) of W/O emulsions composed of (a) soybean oil and (b) hexadecane as oil phase stabilized by (A) PGPR, (-) span 80 and (C) lecithin. Full symbolsare related to 30:70 W:O and open symbols, to 60:40 ratios.

F.Y. Ushikubo, R.L. Cunha / Food Hydrocolloids 34 (2014) 145e153 149orated by the size distribution at lower PGPR concentration, inwhich the rst peak decreases with the lower emulsier concen-tration (data not shown). The second peak, in the range between 1and 3 mm, represents the water droplets of the emulsions, whichmatches the droplet size observed in microscopy images.

Overall, the particle size distribution of wateresoybean oilsystems with Span 80 or lecithin at both W:O ratios were similar,with the rst peakmode in the range of 0.9e1.3 mmand other peaksFig. 7. Optical microscopy images of W/O emulsions produced with soybean oil as oil phasedispersity (Fig. 9b). The mean diameter (d43) matched the dropletsize qualitatively observed in the microscopic images, in whichdroplets of emulsions at 30:70 ratio were larger than those at60:40, and droplets of Span 80 emulsions were smaller than thoseof PGPR ones. In the case of lecithin emulsions, a bimodal distri-bution was observed at both W:O ratios, with one peak mode of5 mm and a second one near 40 mm in the 30:70 ratio emulsion, and13 mm and 91 mm at 60:40 ratio. The bimodal distribution,and different emulsiers at 30:70 and 60:40 W:O ratios. The scale bar is 10 mm length.

hase

F.Y. Ushikubo, R.L. Cunha / Food Hydrocolloids 34 (2014) 145e153150combined with the microscopy images, indicates the occurrence ofocculation and coalescence of the droplets in the emulsions withlecithin.

Fig. 8. Optical microscopy images of W/O emulsions produced with hexadecane as oil p3.5. Rheological essays

The ow curves obtained for the emulsions with soybean oil orhexadecane were tted to the HerscheleBulkley (HB) model andthe rheological parameters are presented in Table 3. The emulsionsproducedwith soybean oil and PGPR showed a Newtonian behaviorat 30:70 W:O ratio. When the water content was increased (60:40ratio), the higher volume of water droplets became susceptible tothe shear deformation, characterizing a shearethinning uid.Besides, an increase in the apparent viscosity was observed.

Furthermore, just after the homogenization, low viscosity uidswere obtained with water and soybean oil with lecithin, withNewtonian behavior at 30:70 ratio and shear-thinning at 60:40W:O ratio. Those uids were unstable and phase separated. In oneday, the systems presented yield stress and higher apparent

Fig. 9. Particle size distribution of W/O emulsions composed of water and (a) soybean oil orrelated to 30:70 and open symbols, to 60:40 W:O ratios.viscosity. That characterizes the formation of a gel network, whichwas intensied along the observation time.

On the other hand, the wateresoybean oil systemwith Span 80,mainly at 60:40W:O ratio, presented yield stress and high apparent

and different emulsiers at 30:70 and 60:40 W:O ratios. The scale bar is 10 mm length.viscosity just after the homogenization. The systems presentedBingham plastic and HB uid behavior at 30:70 and 60:40 W:Oratios, respectively. After one day, at 30:70 ratio, the rheologicalcharacteristics practically did not change with time. At 60:40, theapparent viscosity at 3 s1 decreased, probably due to a change intoa weaker structure.

The waterehexadecane emulsions presented a shearethinningbehavior in most of the conditions, except at 30:70 W:O ratiousing lecithin as emulsier, which showed a Newtonian behavior.The shearethinning behavior indicates that the droplets of waterformed in the hexadecane are affected by the shear deformation.Most of the emulsions presented lower apparent viscosity than thewateresoybean oil systems, since the viscosity of pure hexadecane(0.003 Pa s) is lower than that of soybean oil (0.050 Pa s) and nostructured system was formed. Most of the samples showed an

(b) hexadecane stabilized by (A) PGPR, (-) span 80 and (C) lecithin. Full symbols are

increase in the apparent viscosity after one day because of the oilphase separation. The exceptions were the PGPR and Span 80emulsions at 60:40 W:O ratio, which were the most stable amongthe hexadecane systems.

along time. The relation of the complexviscoelasticmodulus and thestability of the emulsions was discussed in many studies in the

literature (Bouyer et al., 2011; Laza-Knoerr, Huang, Grossiord,Couvreur, & Gref, 2011; Santini, Liggieri, Sacca, Clausse, & Ravera,2007). The good elasticity of the interface could be attributed tothe anchoring of the PGPR hydrophobic chains into the branchedstructure of soybean oil, formed by oleic fatty acids (Benichou,Aserin, & Garti, 2001). In addition, the hydrophilic portion couldalso have played an important role in the anchoringof the emulsiermolecule into the aqueous phase. The hydrophilic portion of PGPR isconstituted by polyglycerol moieties, which are relatively large. Thehydroxyl groups bond to the water molecules by hydrogen bonds,resulting in a structured interface. The good xation of the hydro-philic part of the emulsierwithwater contributed to the stability ofemulsions with higher water volume fraction (60:40 W:O ratio).

On the other hand, the emulsions prepared with hexadecaneand PGPR were not as stable as the soybean oil emulsions. The

Table 2Volume-weighted mean diameter (d43) and span value of the particle size distri-bution of W/O emulsions. Different letters between the samples with same oil andemulsier, at different W:O ratio, indicate signicant difference at p < 0.05.

Emulsier Soybean oil Hexadecane

30:70 60:40 30:70 60:40

d43 (mm) Span d43 (mm) Span d43 (mm) Span d43 (mm) Span

PGPR 1.15a 1.15 1.19a 3.54 5.29a 1.20 4.99b 1.14Span 80 42.54a 1.72 59.55b 3.91 3.74a 0.99 2.86b 1.19Lecithin 35.73a 1.48 73.37b 1.83 38.23a 0.91 82.41b 0.99

; h3:ase)ffere

F.Y. Ushikubo, R.L. Cunha / Food Hydrocolloids 34 (2014) 145e153 151Table 3Rheological parameters (s0: yield stress, Pa; k: consistency index, Pa sn; n: ow indexafter the emulsication and one day after in samples with phase separation (lower phletters between the same sample at different measured times indicate signicant di3.6. Discussion

Among the systems in which emulsions were formed, 60:40W:O ratio emulsions were more stable than 30:70 ones. The mainreason is that, at 60:40 ratio, with larger volume and number ofwater droplets, the apparent viscosity increased, intensifying thehydrodynamic interaction, which led to a lower sedimentation rate.In the same way, a faster destabilization of emulsions with hex-adecane occurred because this oil has lower viscosity than thesoybean oil, which decreases steric stabilization. Besides, the higherinterfacial tension of the waterehexadecane interface and thebigger difference between the water and hexadecane densities ledto a faster phase separation in the emulsions with hexadecane inrelation to those with soybean oil.

The wateresoybean oil emulsions stabilized by PGPR showedhigh kinetic stability. The good emulsifying properties of PGPR inW/O emulsions with triglycerides as the oil phase was reported bya number of studies (Garti, Aserin, Tiunova, & Binyamin, 1999;Mrquez, Medrano, Panizzolo, & Wagner, 2010; Pawlik, Cox, &Norton, 2010; Su, Flanagan, Hemar, & Singh, 2006). The highstability is supported by the small mean droplet size distribution,shown by both microscopic images and particle size distributionmeasurements. Small water droplets are more stable in oil phasebecause of the lower sedimentation speed. The droplet breakupwasfavored by the fast interfacial tension drop, from 22 mN/m (waterepure oil) to 6.7 mN/m (wateresoybean oil with PGPR) (Fig. 4).Furthermore, the coalescence of the droplets was avoided by theviscoelastic characteristic of the interface, shown by the relativelyhigh complex modulus (Fig. 5), which was kept practically constantEmulsier 30:70

s0 k n

Soybean oil PGPR 0 0.16 1Span 80 (t 0) 7.66 0.13 1Span 80 (t 1 day) 9.64 0.12 1Lecithin (t 0) 0 0.16 1Lecithin (t 1 day) 3.00 0.29 0.8

Hexadecane PGPR (t 0) 0 0.07 0.7PGPR (t 1 day) 0 0.69 0.5Span 80 (t 0) 0 0.03 0.8Span 80 (t 1 day) 0 0.10 0.7Lecithin (t 0) 0 0.01 1Lecithin (t 1 day) 0 0.30 0.7lower stability may be related to the linear structure of hexadecane,which did not promote a good xation of the PGPR on the interface.With that, a weaker oileemulsier interaction occurred at theinterface, leading to a faster destabilization.

However, Span 80 was a good emulsier in a system composedof hexadecane at 60:40 W:O ratio. The high stability at high watercontent was supported by the hydrogen bonds formed between thehydroxyls of the hydrophilic moieties of Span 80 (sorbitan) and thewater molecules. In addition, the emulsion at this conditionshowed enhanced stability due to the small droplets, as observed inthe particle size distribution and the images obtained throughoptical microscopy. Those droplets were formed favored by therapid migration of Span 80 to the waterehexadecane interface, asindicated by tension drop between the waterehexadecane inter-facial tension (40 mN/m) to the initial tension with Span 80 in thissystem (6.1 mN/m) (Fig. 4). Regarding the complex modulus,although it was low (Fig. 5), it was constant with time. This indi-cates that the interface was stable evenwith the perturbation of thesystem. This stable elastic network could avoid the droplet coa-lescence. The explanation of the structured interface can be given interms of the hydrophobic chainechain interaction between Span80 and hexadecane, which was more compatible than interactionsbetween PGPR and hexadecane chains. As it is observed in Fig. 3,the hydrophobic part of Span 80 molecule has only one unsatura-tion in its hydrophobic chain, while PGPR hydrophobic componentshave one unsaturation and one hydroxyl (Fig. 2), which hasa hydrophilic character. Since hexadecane is highly hydrophobic,with no unsaturation in its structure, this would create a betterinteraction with Span 80, which has a more hydrophobic chain. Onthe other hand, PGPR showed more compatibility with soybean oil,since this oil is less hydrophobic. The chain compatibility betweensurfactant and oil was described by Chattopadhyay, Shah, andGhaicha (1992).

apparent viscosity at 3 s1, Pa s) of W/O emulsions at two measurement times: just. Soybean oilePGPR emulsions did not show phase separation after one day. Differentnce at p < 0.05.

60:40

h3 s0 k n h3

0.16 0 2.36 0.79 1.251.77a 10.30 4.45 0.79 20.40a

2.06b 24.70 3.49 0.69 8.16b

0.16a 0 0.07 0.77 0.11a

7 1.21b 2.91 0.19 0.83 0.95b

2 0.06a 0 0.30 0.73 0.46a

5 0.60b 0 0.37 0.71 0.46a

2 0.04a 0 0.30 0.67 0.50a

4 0.18b 0 0.33 0.66 0.56b

0.01a 0 0.07 0.79 0.03a1 0.23b 0 0.93 0.57 0.67b

Murdan, S., van den Bergh, B., Gregoriadis, G., & Florence, A. T. (1999). Water-in-sor-

International, 33, 3e14.

HyRegarding the lecithin systems, the kinetic stability was verylow, mainly inwaterehexadecane systems. Themicroscopic imagesof the systems with hexadecane showed the formation of anemulsion with very large and polydisperse water droplets, whichexplains the fast emulsion destabilization. The reason for the big-sized droplets was the slow migration of lecithin molecules to theinterface, which hindered the droplet breakup. The low mobility ofthe emulsier was supported by the short decrease of the interfa-cial tension of the waterepure oil interface (22 and 40 mN/m forwateresoybean oil and waterehexadecane, respectively) to theinterface with lecithin (9.2 mN/m for soybean oil and 24.1 mN/mfor hexadecane) (Fig. 4). The slow mobility of lecithin molecules isprobably due to the presence of charged groups in the hydrophilicmoiety, which have low mobility in the hydrophobic medium. Themobility is even lower in hexadecane, which is more hydrophobicthan soybean oil. Another reason for the low stability of lecithinemulsions is the fact that the lecithin tested in this study did nothave its phospholipids concentration modied. As it is alreadyknown, PC enriched lecithin is better for oil-in-water emulsions,while lecithin with higher concentration of PE and PI are moreappropriated to produce water-in-oil emulsions (Wu & Wang,2003). Furthermore, the aqueous phase separation observed inemulsions with lecithin at high water content (60:40 W:O ratio)(Table 1) suggests the weak interaction of the hydrophilic part ofthe emulsier with water molecules. Lecithin hydrophilic moietyhas only one hydroxyl and ionic groups (Fig. 1), which cannot forma structured interface as suggested in PGPR and Span 80 emulsions.

Moreover, considering the difference in the visual appearanceand in the microscopic images of the Span 80 systems preparedwith soybean oil and hexadecane, it can be seen that distinctstructures were formed. A macroemulsion was formed in thewaterehexadecane system, while a gel-like structure was observedin the wateresoybean oil system. This gelled structure is probablyan organogel formed by self-assembly structures of Span 80molecules in oil in the presence of water. Sorbitan esters, such asSpan 20 (Yu, Shi, Liu, & Huang, 2012), Span 60 (Murdan, van denBergh, Gregoriadis, & Florence, 1999) and Span 80 (Bhattacharya,Kumar, Sagiri, Pal, & Ray, 2012) have been reported to form orga-nogels. Similar results were observed in emulsions prepared withlecithin. The systems with soybean oil presented a structure ofa gel, although with a different aspect from the Span 80esoybeanoil system. The microscopic images showed the formation of manyirregular-shaped aggregates, similarly as observed by Knoth,Scherze, and Muschiolik (2005) in an emulsion prepared withsunower oil and phosphadidylcoline-depleted lecithin. As well asin the Span 80esoybean oil system, those aggregates are thought tobe an auto-organized structure. In fact, lecithin can form a verywide range of structures, such as monomolecular layers, bimolec-ular lms, liposomes and vesicles, liquid crystals, besides emul-sions, microemulsions and organogels (Shchipunov, 1997).According to Vintiloiu and Leroux (2008), in the presence of organicsolvents, reverse micelles are formed, which convert in cylindricalreverse micelles with the addition of a polar solvent. Thosemicellesform a gel network, resulting in a weak organogel.

4. Conclusion

This study showed that different mechanisms act on the stabi-lization of water-in-oil emulsions. The main stability mechanismagainst the sedimentation of a water-in-oil emulsion was the stericstability, which depends on the hydrodynamic interactionsbetween the droplets that, in turn, is dependent on the apparentviscosity. Higher water internal volume resulted in more viscousemulsions, increasing the stability. On the other hand, higher

F.Y. Ushikubo, R.L. Cunha / Food152volume of less viscous oil showed rapid phase separation.Santini, E., Liggieri, L., Sacca, L., Clausse, D., & Ravera, F. (2007). Interfacial rheologyof Span 80 adsorbed layers at parafn oilewater interface and correlation withthe corresponding emulsion properties. Colloids and Surfaces A: Physicochemicaland Engineering Aspects, 309, 270e279.

Shchipunov, Y. A. (1997). Self-organising structures of lecithin. Russian ChemicalReviews, 66, 301e322.

Su, J., Flanagan, J., Hemar, Y., & Singh, H. (2006). Synergistic effects of polyglycerolester of polyricinoleic acid and sodium caseinate on the stabilization ofwater-oil-water emulsions. Food Hydrocolloids, 20, 261e268.

Tadros, T. F. (2008). Applied surfactants: Principles and applications (2nd ed.).Weinheim: Willey-VCH.bitan monostearate organogels. Journal of Pharmaceutical Sciences, 88, 615e619.Pawlik, A., Cox, P. W., & Norton, I. T. (2010). Food grade duplex emulsions designed

and stabilized with different osmotic pressures. Journal of Colloid and InterfaceScience, 352, 59e67.

Poth, U. (2001). Drying oils and related products. In Ullmanns encyclopedia ofindustrial chemistry (pp. 621e636). Weinheim: Wiley-VCH.

Rousseau, D. (2000). Fat crystals and emulsion stability e a review. Food ResearchFurthermore, the rapid decrease in the interfacial tension with theemulsier addition was also important for the formation of smalldroplets, which resulted in higher stability against gravity force.Other factor was the viscoelasticity of the lm at the watereoilinterface, which can prevent the coalescence of droplets. Moreover,depending on the composition of the system, a gel network can beformed, resulting in a product with different rheological charac-teristics and distinct stability mechanisms.

Acknowledgment

The authors would like to thank FAPESP (2009/54.137-1, 2010/16.708-4 and 2011/06.083-0) and CNPq (304611/2009-3) for thenancial support.

References

Benichou, A., Aserin, A., & Garti, N. (2001). Polyols, high pressure, and refractiveindices equalization for improved stability of W/O emulsions for foodapplications. Journal of Dispersion Science and Technology, 22, 269e280.

Bergensthl, B., & Fontell, K. (1983). Phase equilibria in the system soybean lecithin/water. Progress in Colloid & Polymer Science, 68, 48e52.

Bhattacharya, C., Kumar, N., Sagiri, S. S., Pal, K., & Ray, S. S. (2012). Development ofspan 80 e tween 80 based uid-lled organogels as a matrix for drug delivery.Journal of Pharmacy and Bioallied Sciences, 4, 155e163.

Bouyer, E., Mekhlou, G., Le Potier, I., de Kerdaniel, T., du, F., Grossiord, J. L., et al.(2011). Stabilization mechanism of oil-in-water emulsions by beta-lactoglobulinand gum arabic. Journal of Colloid and Interface Science, 354, 467e477.

Bueschelberger, H. G. (2006). Lecithins. In R. J. Whitehurst (Ed.), Emulsiers in foodtechnology (pp. 1e39). Oxford: Blackwell Publishing, (chapter 1).

Chattopadhyay, A. K., Shah, D. O., & Ghaicha, L. (1992). Double-tailed surfactantsand their chain length compatibility in water-in-oil emulsions. Langmuir, 8,27e30.

Claesson, P. M., Blomberg, E., & Poptoshev, E. (2004). Surfaces forces and emulsionstability. In S. E. Friberg, K. Larsson, & J. Sjblom (Eds.), Food emulsions (4th ed.).New York: Marcel Dekker, (chapter 7) (e-book version).

Cottrell, T., & van Peij, J. (2006). Sorbitan esters and polysorbates. In R. J. Whitehurst(Ed.), Emulsiers in food technology (pp. 162e185). Oxford: Blackwell Publishing.

Dickinson, E. (1993). Towards more natural emulsiers. Trends in Food Science &Technology, 4, 330e334.

Garti, N., Aserin, A., Tiunova, I., & Binyamin, H. (1999). Double emulsions ofwater-in-oil-in-water stabilized by a-form microcrystals. Part 1: selection ofemulsiers and fat microcrystalline particles. Journal of the American OilChemists Society, 76, 383e389.

Knoth, A., Scherze, I., & Muschiolik, G. (2005). Stability of water-in-oil emulsionscontaining phosphatidylcoline-depleted lecithin. Food Hydrocolloids, 19, 635e640.

Laza-Knoerr, A., Huang, N., Grossiord, J.-L., Couvreur, P., & Gref, R. (2011). Interfacialrheology as a tool to study the potential of cyclodextrin polymers to stabilize oilewater interfaces. Journal of Inclusion Phenomena and Macrocyclic Chemistry, 69,475e479.

Mrquez, A. L., Medrano, A., Panizzolo, L. A., & Wagner, J. R. (2010). Effect of calciumsalts and surfactant concentration on the stability of water-in-oil (w/o)emulsions prepared with polyglycerol polyricinoleate. Journal of Colloid andInterface Science, 341, 101e108.

McClements, D. J. (2005). Food emulsions: Principles, practices, and techniques (2nded.). Boca Raton: CRC Press (e-book version).

drocolloids 34 (2014) 145e153Vintiloiu, A., & Leroux, J.-C. (2008). Organogels and their use in drug delivery ea review. Journal of Controlled Release, 125, 179e192.

Wendel, A. (2000). Lecithin. In Kirk-Othmer encyclopedia of chemical technology.New York: John Wiley & Sons.

Weyland, M., & Hartel, R. (2008). Emulsiers in confectionery. InG. L. Hasenhuettl, & R. W. Hartel (Eds.), Food emulsiers and their applications(2nd ed.). (pp. 285e305) New York: Springer.

Wu, Y., & Wang, T. (2003). Soybean lecithin fractionation and functionality. Journalof the American Oil Chemists Society, 80, 319e326.

Yu, H., Shi, K., Liu, D., & Huang, Q. (2012). Development of a food-grade orga-nogel with high bioaccessibility and loading of curcuminoids. Food Chemistry,131, 48e54.

F.Y. Ushikubo, R.L. Cunha / Food Hydrocolloids 34 (2014) 145e153 153

Stability mechanisms of liquid water-in-oil emulsions1. Introduction2. Materials and methods2.1. Emulsion preparation2.2. Interfacial tension and dilational rheology2.3. Evaluation of kinetic stability2.4. Optical microscopy2.5. Particle size distribution2.6. Rheological essays2.7. Statistical analysis

3. Results3.1. Interfacial tension3.2. Interfacial dilational rheology3.3. Kinetic stability3.4. Optical microscopy imaging and particle size distribution3.5. Rheological essays3.6. Discussion

4. ConclusionAcknowledgmentReferences