lable at ScienceDirect

Food Hydrocolloids 39 (2014) 151e162

Contents lists avai

Food Hydrocolloids

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

Ultrasonic preparation of stable flax seed oil emulsions in dairysystems e Physicochemical characterization

Akalya Shanmugama, Muthupandian Ashokkumar a,b,*a School of Chemistry, University of Melbourne, Melbourne, Victoria 3010, AustraliabChemistry Department, King Abdulaziz University, Jeddah, Saudi Arabia

a r t i c l e i n f o

Article history:Received 5 November 2013Accepted 6 January 2014

Keywords:UltrasoundEmulsificationNutraceuticalFlax seed oilWhey proteinUltraturrax

* Corresponding author. School of Chemistry, Unbourne, Victoria 3010, Australia. Tel.: þ61 3 83447090

E-mail address: masho@unimelb.edu.au (M. Ashok

0268-005X/$ e see front matter � 2014 Elsevier Ltd.http://dx.doi.org/10.1016/j.foodhyd.2014.01.006

a b s t r a c t

This study reports the incorporation of 7e21% of flax seed oil in pasteurized homogenized skim milk(PHSM) using high intensity ultrasound (US) at 20 kHz between 1 and 8 min and at varying power levels.A minimum process time of 3 min at an applied acoustic power of 176 W was sufficient to produceemulsion droplets (7% oil) with an average mean volume diameter of 0.64 mm and they were stable atleast 9 days at 4 � 2 �C. The mechanical, cavitational and cavitation-after-effects of US are responsible forthe production of smaller sized emulsion droplets and process-induced modifications of milk proteins. Avery small proportion (less than 20%) of partially denatured whey proteins provided stability to theemulsion droplets. The emulsion droplets also showed a surface potential of about �30 mV due to theadsorbed proteins, which provided further stability to the emulsion droplets due to electrostaticrepulsion. In order to see if other high shear techniques can generate stable emulsions, experiments werecarried out using Ultraturrax (UT) at similar energy densities to that of US. UT did not produce stableemulsions until 20 min of processing suggesting the superiority of US emulsification process.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Consuming healthy food has become a major trend in the lastdecade leading to the development of novel food processing tech-niques for the encapsulation and delivery of bioactives/nutraceut-icals. Most bioactives and nutraceuticals are hydrophobiccompounds. A poor water solubility of these compounds causesenormous difficulties in delivering them in food. The delivery ofsuch oil-based bioactives/nutraceuticals as emulsions is well-known (Garti & Yuli-Amar, 2008; Couedelo et al., 2011). Conven-tionally, food emulsions (O/W) are obtained using high shearmixtures such as UT and piston homogenizers with the assistanceof emulsifiers and stabilizers to achieve considerable emulsionstability upon storage (Dapcevic Hadnadev, Dokic, Krstonosic, &Hadnadev, 2013; Maali & Mosavian, 2013; Santana, Perrechil, &Cunha, 2013). The use of large quantities of emulsifiers and Non-GRAS (Generally Recognised As Safe) additives are not permittedin foods, this makes the food industry to rely only on a few range ofemulsifiers and this poses a huge challenge in the area of newproduct development.

iversity of Melbourne, Mel-; fax: þ61 3 93475180.kumar).

All rights reserved.

US has been used for creating emulsions in foods (Soria &Villamiel, 2010; Wulff-Perez, Torcello-Gomez, Galvez-Ruiz, &Martin-Rodriguez, 2009). Much of the existing researchwork in thearea of ultrasonic emulsification has focussed mainly on simplematrix such as an emulsion of sunflower oil inwater. The delivery ofbioactives in a complex/real foodmatrix (health beverage) using USremains as a vast area to be explored. Unlike simple matrices, acomplex food matrix is composed of proteins, carbohydrates, fat,water, vitamins andminerals. Few studies have identified the use ofemulsifiers/surfactants in production of smaller oil droplet (Jafari,He, & Bhandari, 2006; Kentish et al., 2008; Leong, Wooster,Kentish, & Ashokkumar, 2009), however the stability during thestorage of the product has not been studied in detail. In addition,formation of emulsions ultrasonically without the use of externalstabilizers and emulsifiers (food additives) also remains unex-plored. Hence, the purpose of this study is to deliver stable emul-sions of a hydrophobic bioactive compound in a complex foodmatrix such as milk using ultrasound.

In the past, some studies have reported the preparation of soyoil emulsions (O/W) using milk fat globular membrane (MFGM) asan emulsifier and by employing high pressure homogenization(HPH) and microfluidizer (Corredig & Dalgleish, 1998; Roesch,Rincon, & Corredig, 2004). Biasutti, Venir, Marchesini, andInnocente (2010) have produced 15% O/W model dairy emulsionsusing milk cream and emulsifiers by HPH. However, the milk cream

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162152

by itself is rich in MFGM (Harjinder, 2006) and the storage stabilityof these model emulsions was not discussed. In our study, we wereinterested in loading a significant amount of flax seed oil (7e21%) inpasteurized homogenized skim milk (PHSM) without using MFGMand other food additives. The incorporation of flax seed oil in PHSMby employing high intensity US has not been reported in theliterature.

Flax seed oil is a widely popular non polar bioactive among thefunctional food category. It is obtained from flax seed (Linum usi-tatissimum L.) and is a rich source of omega-3-fatty acid, a-linolenicacid (ALA) (Carter, 1993; Mantzioris, James, Gibson, & Cleland,1994). ALA is an essential fatty acid and is known to support cell,nerve & cognitive skills development in children and cardiovascularfunctions in humans (Joshi et al., 2006; Mazza, Pomponi, Janiri,Bria, & Mazza, 2007). In 2002, the Food and Nutrition Board ofthe US (Institute of Medicine) has recommended the adequateintake (AI) levels for ALA in adults (19 years and older): 1.6 g/day formen and 1.1 g/day for women to avoid any deficiency which willresult in symptoms like scaly dermatitis. In addition, Mazza et al.(2007) have indicated a safe and effective dose of flax seed oiland are 3e6 g/day to prevent and treat neurodegenerativedisorders.

Broadly, the aim of our study was to emulsify 7e21% of flax seedoil (59.9% ALA) in PHSM using US at 20 kHz under various experi-mental conditions. The stability of the emulsions was characterisedby a number of techniques and compared with UT emulsificationprocess.

2. Materials and methods

2.1. Materials

Fresh PHSM was purchased from a local supermarket andimmediately stored at 4 �C until further use. The composition of themilk was 3.5% protein, 0.1% fat, and 4.9% lactose as labelled by themanufacturer. The manufacturer’s specification was cross-checkedin our lab for the proteins. The protein content was 3.46% byBradford Assay. We did not observe any creaming-off in the PHSMsample for about 10 days of storage at 4 � 2 �C indicating that fatcontent was very low.

Ultra pure (MilliQ) water was used in all experiments. Unrefinedorganically grown cold pressed flax seed oil with 59.9% of ALAwas agift sample from Stoney Creek Oil Products Pty Ltd, Australia.

2.2. Emulsification by US and UT

The emulsion composition was 7% flax seed oil (v/v) in 93%PHSM, unless mentioned otherwise. Both the water and oil phaseswere added sequentially to the water jacketed glass vessel and thesonicator hornwas positioned at a depth of 0.3� 0.1 cm. Emulsionswere obtained as 50 ml aliquots using a 20 kHz, 450 W ultrasonichorn (12 mm diameter, Branson Sonifier, Model 102 (CE)) at 88, 132and 176 W of nominal applied powers (NAP) for different pro-cessing times from 1 to 8 min. During sonication, thermostatedwater was circulated continuously through a jacket surroundingthe sonication cell and the water temperature was maintained at22.5� 2 �C. The emulsified samples were stored in a refrigerator forabout 9 days at 4 � 2 �C. The analysis and storage studies wereperformed on both fresh and stored samples. In this paper, flax seedoil/milk, flax seed oil/water and oil/water emulsions are referred toas OM, OW and O/W, respectively. The term “unstable emulsion”refers to the system where the oil phase separated within 3 h ofstanding at room temperature whereas “good emulsion” refers tothe system where the oil phase did not separate until 2 days and

“stable emulsion” refers to the system where the oil phase did notseparate until at least 9 days of storage at 4 � 2 �C, respectively.

UT emulsions were prepared in the same jacketed vessel usingUltraturrax (IKA-Labortechnik) at speed dial value 4; 17,500 RPM(22.5 � 0.5 �C) and at an energy density equivalent to that of US(discussed in Results section).

2.3. Particle size and zetapotential measurements

The particle size of the OM emulsionwas measured on fresh andstored samples using a laser diffraction method by Mastersizer2000 (Malvern Instruments Ltd., Worcestershire, U.K). A fewdroplets of the emulsion were suspended directly in recirculatingwater (1250 rpm, obscuration (14e16%) and refractive index of flaxseed oil 1.475). The volume size distribution values viz., D(4,3),Dv90 and Dv50 were recorded. D(4,3) represents volume meandiameter; Dv90 represents the diameter wherein 90% of the vol-ume distribution is below this value; Dv50 represents the diameterwherein 50% of the volume distribution is above and below thisvalue. z-Potential of oil droplets was determined using a ZetasizerNano ZS (Malvern Instruments Ltd., Malvern, Worcestershire, UK).The emulsions were diluted 200 fold using MilliQ water prior tomeasurements. In this paper, D(4, 3) values are mostly used in thediscussion section.

2.4. Creaming stability

The emulsions were visually checked for phase separation andoiling-off or creaming. Also, the amount of creaming was measuredby storing them in 6 ml sealed graduated tubes at 4 � 2 �C for 9days. In this test, Sudan III dye was used to improve the clarityamong separated phases. 0.0025% of Sudan III dye was mixed withflax seed oil for 2.5 h at room temperature using a magnetic stirrer.Instead of flax seed oil, oil-colour mixture was used in making theemulsions. The emulsion stability against creaming was monitoredby measuring the volume of the lipid-rich layer on top (VL) and thevolume of total emulsion (VE) in the tube. Creaming stability interms of creaming index (%) was obtained using the equation (1),

Creaming Index ð%Þ ¼ ðVE � VL=VEÞ � 100 (1)

For example, if the creaming index is 100%, there is no phaseseparation in the emulsions.

2.5. Hydrophobicity of milk proteins

Changes to the hydrophobicity of the milk proteins weremeasured on the aqueous portion of the sonicated emulsions.Aqueous phase of the emulsions were separated by skimming offthe fat at 14,000 rpm for 15 min using Hermle centrifuge (Z306) atroom temperature (Pearce & Kinsella, 1978). Before fluorometricassay, the aqueous phase was vortexed for 30 s and diluted withMilliQ water in the ratio of 1:20 in order to prepare the aqueousmilk protein solution. The changes to the protein content of theaqueous milk protein solution was monitored by the BradfordAssay according to manufacturer’s instructions (Sigma Aldrich PtyLtd, Sydney, Australia) at 595 nm using UV-VIS Spectrophotometer(Carey 3E, Varian, Palo Alto, CA, USA). A stock solution of 0.008M 1-anilinonaphthalene-8-sulfonate (ANS) was prepared in 0.1 M pH 7phosphate buffer. It was wrapped in aluminium foil to prevent lightexposure and stored at room temperature.

In the assay, the required amounts of aqueous milk protein so-lutions were made up with 10 ml of phosphate buffer and 20 mL ofANS solution to obtain a set of diluted samples of aqueous milkprotein solution at different dilution factors viz., N, 65, 33, 22, 16

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

0.1 1 10

Diff

eren

tial V

olum

e Pe

rcen

t (%)

Particle Diameter (μm)

Fig. 1. Volume size distribution of 7% OM emulsions (20 kHz US; 176 W) processed fordifferent times: �1 min, C 2 min, , 3 min, 4 min, > 5 min, * 6 min, 7 min, and8 min on Day 1.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0 1 2 3 4 5 6 7 8 9

Vol

ume

mea

n di

amet

er (

µm)

Sonication Time (min)

Fig. 2. Comparison between volume weighted mean e D(4,3) diameter of flax seed oildroplets in 7% OM emulsions (20 kHz US; 176 W) at different processing times andupon storage at 4 � 2 �C on > Day 1 & C Day 8.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162 153

and 13. By doing this, the protein concentrations of these sampleswere varied indirectly. For example, the dilution factor of 65 and 13refer to 0.00005e0.0002% (w/w) of protein in PHSM and dilutionfactor of N refers to absence of aqueous milk protein solution(Blank). After the addition of ANS, the samples were wrapped in afoil, vortexed immediately and kept in dark for 15 min. The ANSprobe only binds to the hydrophobic regions of the protein that arepresent in sample and fluoresce. Thus in every sample set, the in-crease in the amount of proteins from higher to lower dilutionfactor would lead to increased binding of ANS to the hydrophobicprotein and hence increased relative fluorescence intensity (RFI).The RFI was measured using a spectrofluorimeter (RF-5301PC,Shimadzu, Japan). For hydrophobicity determination using ANS, theexcitation and emission slits and wavelengths were set at 5/5 and390/470 nm, respectively (Chandrapala, Zisu, Palmer, Kentish, &Ashokkumar, 2011). The RFI of each solution was measured start-ing from buffer blank and then from the highest to lowest dilutionof the aqueous milk protein solutions. To obtain Net RFI, RFI of eachdilution blank was subtracted from that of the correspondingprotein solution. Plots of RFI values vs dilution factors of each of theprocessed sample were used to determine the changes to the hy-drophobicity (equal to the slopes) of the milk proteins that werepresent in the aqueous phase of the emulsions.

2.6. Statistical analysis

When necessary, one-way ANOVA with a 95% confidence levelwas used. The ANOVA data with p < 0.05 were considered statis-tically significant.

All the experiments were at least duplicated. The emulsionswere prepared in fresh PHSM samples of the same batch. Allmeasurements were performed on the same day of the sonicationand upon storage, wherever necessary.

3. Results

Fig. 1 shows the volume size distribution of 7% OM emulsionprepared using 20 kHz US at 176Wunder different processing time(Day 1). From the data, it is apparent that an increase in sonicationtime from 1 min to 8 min resulted in a decrease in the size ofemulsion droplets, in particular 3 to 8 min of processing resulted inlarger volume of small sized droplets ranging from 0.1 to 1 mm. Inaddition, the data also highlights the changes to the size distribu-tion pattern, at longer processing times a bimodal distribution isnoted for 1 and 2 min samples, whereas 3e8 min samples showedan increasing tendency towards unimodal size distribution.

Fig. 2 compares the volume weighted mean D(4,3) diameter offlax seed oil droplets in 7% OM emulsions (176 W) at differentprocessing times and upon storage at 4� 2 �C for 8 days. The D(4,3)value shows a 4 fold reduction in size between 1 and 8 min ofprocessing and these values have not changed until 8 days ofstorage.

Fig. 3 shows the creaming stability of 7% OM emulsions (176W&with Sudan (III) dye) upon storage at 4� 2 �C. The creaming index isa measure of the emulsion stability. Emulsions are stable if thecreaming index is 100%. The emulsions processed from 3 to 8 minshow 100% stability against creaming until 9 days of storage, whilethose processed for 1 and 2 min show stability for only 1 day and 2days, respectively. Fig. 4 supports the above statement and is visualevidence showing orange coloured cream layer on the top surfacefor 1 and 2min processed samples against the others on the 9th dayof storage.

Until now, the formation of a stable emulsion was considered tobe the effect of an US process and in order to separate the effect/impact of milk components (in particular, proteins) from the

formation and stability of OM emulsions, the emulsions of milliQwater (no milk protein) and flax seed oil were prepared similar to7% OM emulsions using 176W 20 kHz US between 1 and 8min. Thesize distribution data of flax seed oil/water (OW) emulsions areshown in Fig. 5. From the figure, it is clear that the sizes of the OWdroplets and their distribution are not altered by increasing theprocessing time until 8 min. The distribution tends to remainbimodal and the phase separation occurred immediately after 3 h ofstanding at room temperature (Fig. 6). This observation indicatesthat sonication alone is not sufficient to generate stable emulsionsand milk proteins are important for stabilizing the emulsiondroplets.

90

92

94

96

98

100

102

Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9

Cre

amin

g St

abili

ty %

Storage Time (Days)

Fig. 3. Creaming stability of 7% OM emulsions (20 kHz US; 176 W) processed atdifferent times viz., - 1 min, 2 min and 3e8 min at 176 W upon storage at4 � 2 �C.

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

0.1 1 10

Diff

eren

tial V

olum

e Pe

rcen

t (%)

Particle Diameter (μm)

Fig. 5. Volume size distribution of 7% OW emulsions (20 kHz US; 176 W) processed fordifferent times: 1 min; > 2 min; 3 min; B 4 min; 5 min; þ 6 min; 7 min;8 min.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162154

To confirm the presence of milk proteins on the surface of OMemulsion droplets, the absorbance corresponding to the proteins inaqueous solutions was monitored. The aqueous portion representsthe part of emulsion obtained after centrifugation and dilutionwithmilliQ water (1:20). Fig. 7 compares the absorbance of proteins (at595 nm) present in the aqueous solutions of PHSM and 7% OMemulsions (176 W) processed for 1, 3, 5 and 8 min. The absorbancedue to proteins is lower for all the emulsion samples in comparisonto PHSM.

Fig. 4. Photograph of 7% OM emulsions (20 kHz US; 176 W) with Sudan (III) dyeprocessed for different times on Day 9. First row (1e3): 0 min US (PHSM), 1 and 2 minOM emulsion; Second row (4e6): 3e5 min OM emulsion; Third row (7e9): 6e8 minOM emulsion.

The presence of proteins on the surface of emulsion dropletsmay generate charges on the surface of emulsion droplets. Themagnitude of zetapotential gives an indication of potential stabilityof a colloidal system. Fig. 8 shows the zetapotential values of OMemulsions, PHSM and whole milk on the 9th day of storage(4� 2 �C). All samples were sonicated under similar conditions, viz.,20 kHz 176 W for 1, 5 and 8 min. Though creaming was observed in1 min sample, creamwas removed and the sample was withdrawnafter mixing the uncreamed portion. The 0 min samples of PHSMand whole milk represent the unsonicated (market) milk samples.Within error limits the zetapotential values are around �30 mV forall samples irrespective of their composition and constituents. Theaverage values of zetapotential for whole milk and OM emulsionsare higher than that observed for PHSM.

Fig. 9 shows the changes to fluorescence intensity of aqueousmilk protein solutions obtained from 7% OM emulsions (176 W)processed for 1 min, 3 min and 8 min in comparison to PHSM atdifferent dilution factors. The slope value indicates the changes tofluorescence intensity of each sample at different dilution levels.The changes to the slope values at different sonication time arerelated to changes in the hydrophobicity of proteins present in theaqueous phase. A higher slope corresponds to a greater modifica-tion to the proteins, i.e., an increase in hydrophobicity. Fig. 9 showsthat the fluorescence intensity increases with a decrease in sampledilution (increase in concentration of protein) at any sonicationtime. Thus the hydrophobicity of the samples increases with anincrease in sonication time. The maximum change is noted for3 min sonicated sample.

Table 1 compares the emulsification capacity of 20 kHz US atdifferent NAPs, viz., 88, 132 and 176 W and at different processingtimes from 1 to 8 min (Day 1). The stability of these emulsions waschecked visually. Also, some of them were tested upon storage at4� 2 �C. As already reported,176W produced good emulsions from1 to 8 min of processing; however only those samples obtainedfrom 3 to 8 min of processing remained stable until 8 days (Figs. 1and 2). Processing at 22 W did not produce good and stableemulsions until 8 min of processing (Fig. 10), while processing at132 W produced good and stable emulsions from 3 to 8 min.However, the 3 and 4 min samples creamed-off on 2nd day of

Fig. 6. Photograph of 7% OW emulsions (20 kHz US; 176 W) at 1 and 8 min processingtime and after 3 h of standing at room temperature.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162 155

storagewhile other samples (5e8min) remained stable until 8 days(data not shown). From the volume size distribution (data notshown), a bimodal distribution pattern was observed between 1and 7 min at 88 W,1 and 4 min at 132W and 1 and 2 min at 176W;the tendency towards unimodal distribution increased above theseprocessing times, i.e., until 8 min. Table 1 provides D(4,3), Dv50 andDv90 values of these emulsions. The minimum D(4,3) valuesrequired to produce stable emulsions at 132 and 176Ware 0.53 and0.64 mm at 5 min and 3 min, respectively. Similarly, the minimumDv90 value required to produce stable emulsions at 132 and 176 Ware 1.13 and 1.37 mm at 5 min and 3 min, respectively. Though theD(4,3) and Dv90 values for 8 min 88 W, viz., 0.59 and 1.29 mm areconsidered reasonable and comparable to 5min at 132Wand 3minat 176 W; stable emulsions were not produced in the former.

Fig. 11 represents the volume size distribution of OM emulsionsprocessed by 20 kHz at 176WUS at different oil percentages, viz., 7,15 and 21 on Day 1 and Table 2 shows droplet characteristics of 15and 21% of OM emulsions. This work shows the capability of PHSM(milk) system to hold larger amounts of oil (15 and 21%) in the formof emulsion droplets compared to 7% formulation. To make com-parison between the three different emulsion formulations, theprocess time was kept constant between 3 and 8 min. The stabilityof these emulsions was monitored visually on Day 1. The good and

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

400 500 600 700 800

Abs

orba

nce

(a.u

.)

Wavelength (nm)

Fig. 7. Changes to the absorbance of protein in aqueous phase (1:20 dilution) obtainedfrom the PHSM and 7% OM emulsions (20 kHz US; 176 W) processed for0 min; 1 min, 3 min, 5 min, 8 min at 595 nm(Measurements are average of two repeats).

stable emulsions were produced between 3 and 8 min in case of15% loading and between 6 and 8 min in case of 21% loading. Thedata from Fig. 11 & Table 2 are compared with Fig. 1 & Table 1 andthe key observations are: 1) an increase in oil % resulted in abroadening of size distribution, an increase in the volume of largersized particles and an increase in processing time from 3 to 6min incase of 21% emulsions, 2) D(4,3), Dv50 and Dv90 values of 3 min15% and 6 min 21% emulsions are similar and are 0.85, 0.66 and1.7 mm, respectively, 3) minimum D(4,3) and Dv90 values are notedamongst 8 min emulsion samples and 4) an increase in oil %resulted in a lowering of the efficiency of US to breakup the largeremulsion droplets, i.e., Dv90. The 15 and 21% loading emulsionswere stable until 5 days of storage without any visual creaming at4 � 2 �C.

Table 3 shows the process time necessary to prepare 7% OMemulsions using UT when operated at equivalent energy density of176W 20 kHz US at 1, 2, 5, 7 and 8min. The energy density of 50ml,176 W 20 kHz US emulsions at different process times was calcu-lated using Equation (2),

Energy Density ðJ=mlÞ ¼ ðPower Drawn ðNAPÞ ðWÞ�TimeðsÞÞ=Volume ðmlÞ

(2)

The process times for UT emulsions were obtained usingEquation (2) and by substituting the values of calculated energydensity (US), volume of solution (50 ml) and power drawn by UT ata speed dial value of 4 (17,500 RPM; 70W). The UT processing timeswere 2 min 30 s, 5 min, 12 min 30 s, 17 min 30 s and 20 min incomparison to 1, 2, 5, 7 and 8 min of 176 W 20 kHz US operation.The UT emulsions were characterized and the data are shown inFig. 12 and Table 4. From Fig. 12, it is apparent that all the emulsionsamples made with UT showed bimodal curve with huge amountsof larger sized particles. Also, D(4,3), Dv90 and Dv50 sizes of the UTemulsions are larger compared to those off US emulsions. Forexample, the Dv90 value ranges between 3.1e7.6 mm for UT

-40

-35

-30

-25

-20

-15

-10

-5

00 1 5 8

ζpo

tent

ial (

mV

)

Sonication Time (min)

Fig. 8. Zetapotential values of storage samples representing PHSM, whole milkand - 7% OM emulsions prepared by sonication (20 kHz; 176 W US) from 0, 1, 5 and8 min on 9th day at 4 � 2 �C.

y = 2E+06xR² = 0.9917

y = 3E+06xR² = 0.999

y = 6E+06xR² = 0.9924

y = 7E+06xR² = 0.9966

0

200

400

600

800

1000

1200

Rel

ativ

e Fl

uore

scen

ce in

tens

ity

Dilution Factor of 1:20 Aqueous milk protein solution

-----------------------------------------------------------------------------------------------∞ 65 33 22 16 13

Fig. 9. Changes to the fluorescence intensity of aqueous milk protein solutions (1:20)obtained from 7% OM emulsions (20 kHz US; 176 W) processed for 1 min, 3 minand � 8 min in comparison to PHSM at different dilutions are indicated by their slopevalues.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162156

samples in comparison to 0.75e3 mm for US samples. In addition,Fig. 13 shows emulsion instability of the freshly prepared UT sam-ples. It is seen as a creamy layer on the surface of all the emulsionsamples just within few hours of standing at room temperature.

4. Discussion

The results shown in Figs. 1e10 can be summarized as below(summary points 4.1e4.8):

4.1 The size of OM emulsions decreases with an increase insonication time; higher volumes of small sized droplets areobtained at longer sonication times. A 4-fold reduction ofD(4,3) value is observed at 8 min sonication.

4.2 A bimodal distribution is noted until 2 min of sonication anda tendency towards unimodal distribution is observed from 3to 8 min sonication for OM emulsions.

4.3 The stability of OM emulsions is supported by the D(4,3)values (8 Days; 4 � 2 �C).

4.4 The creaming index of emulsions processed between 3 and8 min is 100% on the 9th day at 4 � 2 �C, while those pro-cessed for 1 and 2 min showed 96 and 99%, respectively.

4.5 The US emulsions prepared using 7% flax seed oil and watershow a bimodal size distribution and are unstable. An in-crease in process time (1e8 min) does not cause reduction inthe size of emulsion droplets.

4.6 The amount of protein present in the aqueous phase of OMemulsions is decreased.

Table 1Emulsification capacity of 20 kHz US to produce 7% OM emulsion at different powerlevels, i.e., 88, 132 and 176 W at different times of processing from 1 to 8 min.

US processed sample at20 kHz and different time and NAP

D(4,3) (mm)a Dv50 (mm)a Dv90 (mm)a

5 min; 88 W 0.84 0.55 1.908 min; 88 W 0.59 0.38 1.293 min; 132 W 0.75 0.47 1.744 min; 132 W 0.70 0.46 1.565 min; 132 W 0.53 0.36 1.138 min; 132 W 0.39 0.29 0.773 min; 176 W 0.64 0.45 1.374 min; 176 W 0.54 0.38 1.145 min; 176 W 0.48 0.34 0.998 min; 176 W 0.40 0.30 0.79

a The results are average of two measurements and the error value varies from0 to 0.017 mm.

4.7 Within experimental errors, the zetapotential values of 7%OM emulsion droplets lies around �30 mV on the 9th day ofstorage at 4 � 2 �C.

4.8 The slope of the relative fluorescence intensity (RFI), i.e.,hydrophobicity increases with an increase in sonication timeand a maximum change is noted for 3 min OM sample.

Stability of 7% OM emulsions at 176 W 20 kHz US

A 7% loading of the flax seed oil chosenwas based on FDA’s GRASnotification number: GRN000256, which recommends the level ofhigh linolenic acid flax seed oil in milk products. The stability of USin making 7% OM emulsions at 176 W are discussed based on par-ticle characterization, creaming & storage stability and by under-standing changes to the components of the milk which may impartstability to these emulsions (summary points 4.1e4.4).

The US processing is considered as a high energy emulsificationtechnique and the emulsions of smaller droplet size are producedby the physical effects generated in liquids, viz., mechanical vi-brations, acoustic streaming, acoustic cavitation, microstreaming,shear and turbulence (Abbas, Hayat, Karangwa, Bashari, & Zhang,2013; Ashokkumar et al., 2010). Canselier et al. (2002) havedescribed US emulsification as a two-step process: in the first step,the turbulence caused by the mechanical vibration leads to theeruption of dispersed phase droplets into the continuous phase andthe second step consists of breaking up of droplets through theshear forces generated by cavitation at the interface. The emulsi-fication process can also be described as two opposite “elementarysteps”: droplet breakup leading to formation of several smallerdroplets, and dropletedroplet coalescence leading to formation of alarger droplet. In general, the droplet-size distribution obtainedduring emulsification is governed by the competition betweenthese two opposite processes (Vankova et al., 2007). The emulsionstability is closely related to the droplet size distribution, sincelarger droplet size distributions may enhance the Oswald ripening,i.e., increasing the size to larger droplets in turn favours dropletcoalescence and creaming (Gutierrez, Rayner, & Dejmek, 2009).

The amount of US energy incorporated into emulsions not onlybreaks the planar interface but also overcomes the Laplace pressurein order to produce finer droplets (Canselier et al., 2002). Laplacepressure (PL) is the pressure difference between the convex and theconcave side of a curved interface of an emulsion droplet and isgiven by PL¼ g (l/R1þ l/R2), where g is the interfacial tension and R1and R2 are the principle radii of curvature. For a spherical drop ofradius r we thus have PL ¼ 2 g/r (Walstra, 1993). Hence, Laplacepressure is dependent on interfacial tension of the droplet andtherefore larger amounts of shear (physical effects of US) arerequired to overcome the interfacial tension in between liquids.Thus in our system, a) the mechanical forces of US helps in themixing of the two immiscible phases and formation of largeremulsion droplets by breaking the initial planar interface and b) theshear forces generated by cavitation helps in counteracting theLaplace pressure to generate finer droplets. However at this stage,the effect of milk components in lowering the interfacial tension (g)cannot be neglected because this would also lower the Laplacepressure and hence lower the amount of shear required to breakthe droplets.

From Figs. 1 and 2, a significant change in size distributionpattern and reduction in D(4,3) value, i.e., 1.38e0.40 mm is observedbetween 1 and 8 min of processing. Here, as the processing timeincreases, the amount of input energy increases as well, leading todisruption of more number of droplets and therefore, a decrease inaverage size of emulsion droplets and possibly the changesobserved in the size distribution pattern (summary points 4.1 &4.2). In a previous study on emulsification by homogenization, the

Fig. 10. Photograph represents the difference between 7% OM emulsions produced by 20 kHz US at 88 and 132 W (as freshly prepared). First row (Left to right shows good andstable emulsions): 5 min and 8 min emulsion at 132 W; Second row (Left to right shows unstable emulsions-as yellowness on top surface): 5 min and 8 min emulsion at 88 W.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162 157

production of smaller droplets was shown to increase by increasingthe duration of homogenization (McClements, Decker, & Weiss,2007).

In addition to droplet size characteristics, the physical stability isan important property of emulsion. An emulsion is physically stableif its dispersed state does not change, i.e., if its droplet size distri-bution remains constant for a particular period of time, forexample, 7 days at chill or refrigerated storage (4 �C) in case ofmarket milk products. From Fig. 2 & summary point 4.3, the volumeweighted mean diameter of 7% OM emulsions did not vary until 8days at 4 � 2 �C. The physical stability is generally achieved bypreventing the droplets from sedimentation (gravitational separa-tion), aggregation or coalescence or Ostwald ripening. It can beachieved by employing suitable emulsifiers or by producingemulsions of similar droplet sizes (Schubert & Engel, 2004).

In our experiments, the creaming index value is 100% for 3e8 min emulsions against 96 and 99%, respectively, for 1 and 2 minemulsions (Figs. 3 and 4 & Summary point 4.4). The creaming indexvalues indicate the gravitational separation of emulsions in theform of creaming. Creaming, the upwardmovement of an emulsiondroplet is hindered in the case of 3e8 min samples. Possible rea-sons include decrease in particle size, increase in repulsion be-tween the electrical charges on the surface of emulsion droplets,etc. (Basaran, Demetriades, & McClements, 1998; Chanamai &McClements, 2000). However, most importantly, the componentsof PHSM can also provide stability to these emulsions. PHSM ismajorly composed of proteins, lactose, fat, vitamins and minerals.Among these, the milk fat globular membrane (MFGM) present onthe surface of milk fat globules and themilk proteins are consideredas the important factors contributing to stability of market milkagainst creaming. Here, the stability of milk fat globule is impartedby different conditions of milk processing, in particular the heatingprocess and homogenization process (pressure) contributes to

different functionality of MFGM along with their associated pro-teins (Aiqian, Singh, Taylor, & Anema, 2002; Cano-Ruiz & Richter,1997). However in PHSM, the stabilized fat globules are present inminor amounts (<0.1%) along with the modified MFGM and in theOM emulsions the fat source (flax seed oil) is naturally devoid ofMFGM, which leads to the importance of milk proteins in the sta-bilization of the emulsion droplets.

The milk proteins are of two major types, viz., caseins and wheyproteins. The isolates of milk caseins and whey proteins obtainedfrom various processing techniques are used as emulsifiers in thefood industry. So these proteins may possibly play a vital role in thestability of 7% OM emulsions generated by US, as reviewed byDickinson, 2001. To confirm this, the emulsion stability of OWemulsions (devoid of PHSM) was monitored in the current work.The important observations are indicated in summary point 4.5: inthe absences of PHSM (milk components), the droplet size ofemulsions and their distribution remained unchanged from 1 to8min of processing (Fig. 5). Also, none of the emulsions were stableand the phase separation occurred immediately after 3 h ofstanding at room temperature (Fig. 6). This is mainly because ofparticleeparticle aggregation (coalescence) and is well-indicatedby its bimodal curve (Kuhn & Cunha, 2012). Hence, the impor-tance of milk components, in particular, the proteins in the makingof stable 7% OM emulsion can be realized as discussed below.

From Figs. 1 and 5, the distribution patterns of 1 and 2 min 7%OW emulsions are similar to those prepared at 1 and 2 min 7% OMemulsions (bimodal) but the D(4,3) sizes of 7% OW emulsion are2.77 and 2.02 mm (Data not shown) in comparison to 1.38 and0.83 mm of 7% OM emulsion, respectively (Table 4). Thus, thepresence of minimal amount of surfactant, such as a protein, isrequired to achieve a reduction in the droplet size of emulsions(Tcholakova, Denkov, & Danner, 2004). Upon processing from 1 to8 min, the Dv90 and D(4,3) values of OWemulsions decreased from

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

0.1 1 10

Diff

eren

tial V

olum

e Pe

rcen

t (%)

Particle Diameter (μm)

Fig. 11. Volume size distribution of emulsion processed by 20 kHz, 176 W US atdifferent oil percentages 1) 15% OM emulsions processed for A 3 min, þ 8 min; and 2)21% OM emulsions processed for 6 min, : 8 min.

Table 2Emulsion characterization of OM emulsions produced by 176 W, 20 kHz US atdifferent oil percentages viz., 15 and 21% between 3 and 8 min.

Amount of flaxseed oil inemulsion (%) v/v

Process timeat 176 W; 20kHz (min)

D(4,3) (mm)a Dv50 (mm)a Dv90 (mm)a

15 3 0.85 0.67 1.7415 4 0.75 0.59 1.5315 5 0.69 0.52 1.4115 8 0.61 0.46 1.2121 6 0.85 0.67 1.7221 7 0.87 0.67 1.7521 8 0.82 0.63 1.66

a The results are average of two measurements and the error value varies from0 to 0.20 mm.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162158

5.54 to 3.72 and 2.77 to 1.87 mm, while the sizes of OM emulsionsreduced from 3.02 to 0.79 and 1.38 to 0.40 mm, respectively. So, asignificant and larger size reduction is noted in OM emulsions incomparison OWemulsions at any processing time. This may be dueto adsorption of milk proteins on the surface of oil droplets, whichcontribute to the lowering of interfacial tension and the Laplacepressure as discussed earlier. Though generic, the adsorption ofprotein at the surface of the emulsion droplets can be confirmedfrom the data shown in Fig. 7. It is indicated by a decrease in theabsorbance of proteins present in the aqueous phase in comparisonto unsonicated PHSM. Therefore, certain amount of protein that

Table 3UT processing of 7% (OM) emulsions at equivalent energy density of 176 We20 kHz US

Sample no Time of USprocessing(min)

Volume in USprocessingcell (ml)

Power drawnby 20 kHzUS (W)

Enerin th(J/m

1 1 50 176 2112 2 50 176 4223 5 50 176 10564 7 50 176 14785 8 50 176 1689

was present in the aqueous phase of emulsions has decreased uponsonication due to their adsorption on the surface of emulsiondroplets (Summary point. 4.6). However, amongst these OMemulsions, creaming was observed with 1 and 2min samples whenthey are stored at 4� 2 �C (Fig. 3). This can be related to incompletecoverage of proteins on the surface of oil droplets at shorter pro-cessing times (Jafari, He, & Bhandari, 2007). The samples processedbetween 3 and 8 min showed creaming stability until 9 days at4 � 2 �C (Fig. 3). Thus in case of 7% OM emulsions, a minimumprocessing time of 3 min is recommended for the coverage of suf-ficient amount of proteins on droplet surface.

The milk protein contribution to the size distribution pattern of7% OM emulsions can be explained by two-phase adsorption ki-netics as discussed by Romero et al. (2011). The first phase ischaracterized by a rapid decrease of interfacial tension and it cor-responds to the phase of protein adsorption. The second phase isattained after a certain amount of time, characterized by a slowevolution of interfacial tension and it corresponds to the phase ofconformational rearrangements of proteins at the O/W interface.The OM samples that are obtained between 3 and 8 min of USprocess had enough time to pass through all phases of adsorptionkinetics and remained stable upon storage unlike the 1 and 2 minsamples (Fig. 3). In order to confirm this, charge on the surface ofemulsion droplets and protein modifications were monitored(Summary point 4.7 & 4.8).

The emulsion droplets often have an electrical charge, whichplays an important role in their functional performance and sta-bility. From Fig. 8, within experimental errors, the zetapotentialvalues were �30 mV for all OM emulsions until 9th day of storage.This net negative charge suggests the presence anionic moleculeson the surface of these emulsion droplets, e.g., caseins, beta-lactoglobulin, etc. (Matsumiya, Takahashi, Inoue, & Matsumura,2010). A value of �25 to �30 mV is enough to create high energybarrier between emulsion droplets which in turn would providegood colloidal stability (Mirhosseini, 2010; Mohammadzadeh,Koocheki, Kadkhodaee, & Razavi, 2013; Mora-Huertas, Fessi, &Elaissari, 2010). Sarkar, Horne, & Singh, 2010, have shown a zeta-potential of �30 mV for emulsions obtained using 20% soybean oiland pure beta-lactoglobulin solution (1%) at pH 7.5. In our emulsionsamples, the pH values were between 6.68 and 6.74 at all pro-cessing times (Data not provided). The good electrostatic repulsiondeduced from the high zetapotential at these pH ranges suggests abetter interfacial packing of proteins on the surfaces of the OMemulsion droplets.

Here, it is also worthwhile to discuss the values of zetapotentialfor both unsonicated and sonicated samples of whole milk andPHSM (1e8 min). They did not change upon sonication of thesample and remained the same within the experimental errors(Fig. 8). As mentioned earlier, during conventional processing, theheat and homogenization conditions have already modified andcreated the interfacial layers of the fat globules present in themarket milk and contributed to their stability (Wade & Beattie,1997). It was mainly due to the association of plasma proteinswith the MFGM of the fat globules and likewise in our work,

at 1, 2, 5, 7, 8 min of processing.

gy densitye emulsionl)

Powerdrawn byspeed dial 4 of UT (W)

Time of UT processing at sameenergy density of US, bycalculation (min:s)

.2 70 02:30

.4 70 05:00

.0 70 12:30

.4 70 17:30

.6 70 20:00

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

0.1 1 10 100

Diff

eren

tial V

olum

e Pe

rcen

t (%)

Particle Diameter (μm)

Fig. 12. Volume size distribution of freshly prepared 7% OM emulsions at processingtime A 2.5 min, 5 min, C 12.5 min, þ 17.5 min, and > 20 min using UT at theequivalent energy density of US (176 W; 20 kHz at 1, 2, 5, 7 and 8 min).

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162 159

sonication is considered to be cause for the presence/association ofplasma proteins, viz., casein or whey proteins of the PHSM on thesurface of flax seed oil droplets. In conclusion, it is clear fromsummary point 4.7 that the adsorbed protein molecules on thesurfaces of oil droplets prevented the flocculation of these oildroplets possibly via electrostatic repulsion as was indicated by Bos& van Vliet, 2001; Tcholakova, Denkov, Ivanov, & Campbell, 2002.

Additionally, the above discussion can also be supported byanalysing the changes to hydrophobicity of milk proteins. Forexample, the emulsifying capacity of heat denatured proteins iscorrelated to surface hydrophobicity index (Nakai, 1983). The in-crease in hydrophobicity is due to the unfolding of the proteinmolecule to expose the hydrophobic residues, thus enhancing theiradsorption at the oilewater interface. During sonication, the in-crease in the process time from 1 to 8 min resulted in an increase ofthe hydrophobicity of the aqueous milk protein solutions (Fig. 9 &Summary point 4.8). Therefore, emulsification by US has affectedthe proteins of the milk that are inherently present in the PHSM.But these changes are not vigorous enough to make themcompletely lipophilic which would rather decrease their emulsifi-cation properties. Hence, during protein modification, a balancebetween hydrophilic and lipophilic groups is ensured (moderate)and it contributed to the emulsion stability of our samples. Asmentioned by Nakai, Cheung, and Voutsinas (1983), the electriccharge and hydrophobicity of the adsorbed proteins play animportant role in the emulsification process. Our previous work hasreported similar modifications to milk proteins, when PHSM wassonicated. It has explained the shear induced denaturation & ag-gregation of whey proteins caused by the phenomenon of acousticcavitation in liquids especially at bubble liquid interfaces. Less than20% of whey proteins were denatured when the PHSM was soni-cated until 8 min at 176 W (Shanmugam, Chandrapala, &Ashokkumar, 2012). Similarly, Chandrapala et al. (2011) have re-ported the changes to the surface hydrophobicity of whey proteinconcentrate (WPC) solutions (4% protein) during sonication. Theyhave shown an increase in surface hydrophobicity, when thesamples were sonicated for 5 min by 20 kHz US. HoweverChandrapala, Martin, Zisu, Kentish, and Ashokkumar (2012) andShanmugam et al. (2012) have ascertained the process stability ofcaseins micelles upon sonication of milk. Pertaining to the above

discussions, it is noted that the whey proteins of the milk are highlysusceptible to high intensity sonication process (20 kHz) in com-parison to the micellar caseins of the PHSM. Hence, these processinduced modifications to the whey proteins are helpful in the for-mation of stable emulsions of flax seed oil and PHSM. Thus, cavi-tation bubble liquid interfaces and the shear forces generated bythe high intensity US can contribute to the functionality of wheyproteins in the PHSM.

In summary, the factors responsible for the formation of stableOM emulsions are: 1) mechanical vibration and acoustic cavitationof US are responsible for themixing of two immiscible phases at theoil and milk interface leading to emulsion formation and 2) theshear forces generated by the US are responsible for the partialdenaturation, increased hydrophobicity of the whey proteins(surfactant characteristics) and decreased emulsion droplet sizeresulting in stability of the emulsions.

Effect of nominal applied power, oil % and processing tech-niques on the stability of OM emulsions

The results shown in Figs.11e13 and Tables 1e4 are summarizedbelow (summary points 4.9e4.16):

4.9 The emulsions processed by 20 kHz US from 1 to 7 min at88 W; 1 to 4 min at 132 W; 1 to 2 min at 176 W are unstablewith bimodal distribution pattern except 8min 88W sample.

4.10 The tendency towards unimodal distribution is observed for5e8 min samples at 132 W and 3e8 min samples at 176 Wand are stable until 9 days at 4 � 2 �C

4.11 Though the values D(4,3) and Dv90 of 8 min 88 W sample iscomparable with 5 min 132Wand 3 min 176W samples, theemulsions are not stable.

4.12 The emulsions at higher oil percentages, viz., 15 and 21% arestable for 5 days at 4 � 2 �C

4.13 Good and stable emulsions are obtained between 3 and8 min with 15% emulsions and 6 and 8 min with 21%emulsions.

4.14 D(4,3), Dv50 and Dv90 values of 3 min 15% and 6 min 21%emulsions are similar.

4.15 UT process times are calculated at the equivalent energydensity of 50 ml, 20 kHz 176 W (NAP) US.

4.16 UT emulsions are bimodal and unstable. The Dv90 values ofthe UT samples are 2.5e4 times larger than the US emulsionsamples.

From summary points 4.9 to 4.11, at constant process times,increasing the sonication power results in a decrease in emulsiondroplet diameter (Table 1). Also, a notable change in the size dis-tribution can be observed. A similar study by Madadlou, Mousavi,Emam-Djomeh, Ehsani, and Sheehan (2009) has reported ahigher breakage of re-assembled casein micelles at higher soni-cation power. In general, at lower frequencies, the bubbles gener-ated in the sound field are relatively large in size and when theacoustic power is increased, the size and the number of cavitationbubbles increase (Ashokkumar & Mason, 2007; Brotchie, Grieser, &Ashokkumar, 2009; Lee, 2005) followed by the intense collapse ofthese bubbles resulting in high shear forces and stronger shock-waves in liquids (Leong, Ashokkumar, & Kentish, 2011) and pro-duction of finer emulsion droplets in our system.

The sonication power (NAP) can also be represented as powerdensity, P (W/ml). The power density, the average energy dissipatedper unit time and unit volume, is a measure of strength of turbu-lence (shear) in solution and the maximum diameter of theemulsion droplet in any turbulent flow is given by Equation (3)(Walstra, 1993):

Table 4Comparison of D(4,3), Dv50 and Dv90 value of US and UT process at same energy density.

US process time20 kHz; 176 W (min:s)

Equivalent UT processtime (min:s)

D(4,3) of US(mm)a

Dv50 of US(mm)a

Dv90 of US(mm)a

D(4,3) of UT(mm)a

Dv50 of UT(mm)a

Dv90 of UT(mm)a

1:00 2:30 1.38 1.12 3.02 3.30 2.44 7.582:00 5:00 0.83 0.59 1.81 2.27 1.75 4.935:00 12:30 0.48 0.34 0.99 1.54 1.30 3.257:00 17:30 0.39 0.30 0.75 2.20 1.30 3.318:00 20:00 0.40 0.30 0.79 1.48 1.23 3.15

a The results are average of two measurements and the error value varies from 0 to 0.428 mm.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162160

dmax ¼ CP�2=5g3=5r�1=5c (3)

where, C is a constant, P is the power density, g is the interfacialtension and rc is the density of the continuous phase. Hence, theshearing effect and ultimately the dmax are affected by the powerdensity. By calculation, the power density is lower for the emul-sions obtained at 88 W in comparison to 132 and 176 W. FromTable 1, it is apparent at any particular sonication time that thedroplets produced at lowest power (88 W) are larger compared tothose produced at 132 and 176 W. Therefore, lower the powerdensity, weaker is the shear generated. In addition to power den-sity, dmax is also dependant on the residence time during the for-mation of emulsions (Karbstein & Schubert, 1995). The processcarried out at 132 W and at a residence time of 8 min has producedsmaller droplets in comparison to 132 W emulsions prepared at aresidence time of 5 min. This shows the importance of optimumcombination of power density and residence time for the formationof smaller emulsion droplets. From Table 1, though the D(4,3) andDv90 values for 8 min 88 W are comparable to 5 min 132 W and3 min 176 W, stable emulsions were not obtained for the followingreason: a combination of residence time and power density wassufficient enough to produce smaller droplet size, but not strongenough to modify the whey proteins.

The preparation of emulsions at higher power levels can inducecoalescence among emulsion droplets and results in the formationof larger emulsion droplets. However, coalescence (increase inD(4,3)), is not observed in our system even at the highest power

Fig. 13. Photograph represents 7% OM emulsions obtained by UT process after fewhours of standing at room temperature. Second row (Left to right): 2 min 30 s, 5 minand 12 min 30 s samples, First row (Left to right): 17 min 30 s and 20 min Samples.Creaming is visible on the surface of all the samples.

level of operation (176 W). These results are similar to the obser-vation recorded by Kentish et al. (2008) who reported an increasein the droplet sizes of 15% OW emulsions (with emulsifiers) onlyabove 200 W NAP.

Good and stable emulsions of 15 and 21% oil were obtained onlybetween 3e8 min and 6e8 min of processing, while the rest of thesamples remained unstable. The instability is possibly due to thepresence large emulsion droplets (Fig. 11) in comparison to the datashown in Fig. 1. It was visually observed in the form of creaming ofsamples and is primarily due to the phenomenon of gravity sepa-ration. It can also be due to the presence of partially covered proteinsurfaces on these larger emulsion droplets. These incompletelycovered surfaces can lead to droplet coalescence. While comparingthe data shown in Figs. 11 and 1, the distribution pattern of all highoil emulsions overlaps with the 2 min 7% 176 W OM emulsion butthe former samples are stable for 5 days unlike the later stable onlyfor 2 days. This effect may possibly be due to the optimum com-bination of residence time and power density in the formercompared to the latter. Here, it is also important to note that the 7%OM emulsion had higher amounts of milk protein in comparison to15 and 21% high oil emulsions. The 7% OM emulsions had about 3%(w/w) of milk protein in the formulation in comparison to about 1%(w/w) of milk proteins in the 21% OM emulsions.

In the following section we again re-emphasise the importanceof cavitation and shear that are produced by the US system againsta sole shear system (UT). In brief, the emulsification capacity of UTat equivalent energy densities of 176 We20 kHz US was evaluated(Summary point 4.15 & 4.16). This work reflects the efficiency of theshear forces that are created by UT process. It is deduced bymeasuring the particle size and by monitoring the stability ofemulsions. In order to produce UT emulsions at the same energydensities of 176 W 20 kHz US, it has to be operated at 2.5 times theprocess time of US, for example 1 min of US process is equivalent to2 min:30 s of UT process (Table 3). From Fig. 12 & Table 4, the sizesof UT emulsions are always higher than US emulsions and the sizedistribution remained bimodal for all UT processes. The emulsioninstability is noted in all the fresh UT samples (Fig. 13). As alreadydiscussed, the emulsion stability indirectly refers to the processinduced modifications on the whey protein and their efficiency inlowering of the interfacial tension to create finer emulsions. Hence,it can be suggested from the above discussion that the shear forcesof UT are not effective to produce good and stable emulsions.Therefore, UT process is not seen as an efficient technique in thegeneration of finer and stable OM emulsions.

5. Conclusions

Stable emulsions of 7% flax seed oil and PHSM were obtainedusing 176 W 20 kHz US. The mechanical, cavitation and cavitationafter-effects of US are responsible for producing stable dairyemulsions for a minimum of 9 days at 4 � 2 �C. The mechanismresponsible for the stability of the emulsions was evaluated bymeasuring the creaming index, zetapotential values and thechanges to the hydrophobicity. The milk proteins, viz., whey

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162 161

proteins have contributed to the stability of these emulsions. Aminimum process time of 3 min is recommended to produce stableemulsions of 7% flax seed oil and milk. The increase in powerdensity or NAP resulted in the production of large number ofsmaller emulsion droplets until 8 min of processing. The impor-tance of residence time and power density was studied, and anoptimum combination of these two was identified to produce sta-ble emulsions. A combination of 5 min & 132W and 3 min & 176Wproduced stable emulsions. Also in this work, emulsions with largequantities of oil, viz., 15 and 21% were obtained using 176W 20 kHzUS. Minimum process times of 3 min and 6 min are recommendedtomanufacture 15 and 21% high oil emulsions at 176W. In addition,this work compares two different homogenization techniques, viz.,US and UT in the preparation of stable 7% OM emulsions. Atequivalent energy densities, US process is efficient in the produc-tion of stable emulsions while UT did not produce emulsions at anyprocessing time. The data discussed suggest that flax seed oil, anon-polar bioactive, can be effectively incorporated into a complexfood matrix in the form of an emulsion by employing US technique.In the formulations, the amount of flax seed oil was varied from 7 to21%, which may provide 4.2 to 12.6% of ALA in the diet. The ALA offlax seed oil are rich source of u-3 fatty acids and this can substitutefish oil, used widely in the preparation of u-3 food/infant formu-lations. Sensorially, the fishy flavours are not detected in the flaxseed oil/milk US emulsions unlike many of the u-3 food formula-tions, which weremanufactured using the fish oil. As flax seed oil isan abundant source of poly unsaturated fatty acids (PUFA), theproduction of rancid flavours are expected upon any processingtechnique. However, in all our emulsion samples the rancid flavouris not noted by the human senses (sensorially). The current paperwill remain as the base to a following paper on the functionalcharacteristics/properties of OM emulsions.

Acknowledgements

The author would like to acknowledge the University of Mel-bourne for providing Melbourne International Fee Remission andMelbourne International Research scholarships and also would liketo thank Stoney Creek Oil Products Pty Ltd, Australia for providingflaxseed oil as a gift sample for research.

References

Abbas, S., Hayat, K., Karangwa, E., Bashari, M., & Zhang, X. (2013). An overview ofultrasound-assisted food-grade nanoemulsions. Food Engineering Reviews, 5,139e157.

Aiqian, Y., Singh, H., Taylor, M. W., & Anema, S. (2002). Characterization of proteincomponents of natural and heat-treated milk fat globule membranes. Interna-tional Dairy Journal, 12, 393e402.

Ashokkumar, M., Bhaskaracharya, R., Kentish, S., Lee, J., Palmer, M., & Zisu, B. (2010).The ultrasonic processing of dairy products e an overview. Dairy Science &Technology, 90, 147e168.

Ashokkumar, M., & Mason, T. J. (2007). Sonochemistry. Kirk-Othmer encyclopedia ofchemical technology (pp. 1e34). John Wiley & Sons, Inc.

Basaran, T. K., Demetriades, K., & McClements, D. J. (1998). Ultrasonic imaging ofgravitational separation in emulsions. Colloids and Surfaces A: Physicochemicaland Engineering Aspects, 136, 169e181.

Biasutti, M., Venir, E., Marchesini, G., & Innocente, N. (2010). Rheological propertiesof model dairy emulsions as affected by high pressure homogenization. Inno-vative Food Science & Emerging Technologies, 11, 580e586.

Bos, M. A., & van Vliet, T. (2001). Interfacial rheological properties of adsorbedprotein layers and surfactants: a review. Advances in Colloid and Interface Sci-ence, 91, 437e471.

Brotchie, A., Grieser, F., & Ashokkumar, M. (2009). Effect of power and frequency onbubble-size distributions in acoustic cavitation. Physical Review Letters, 102,084302.

Cano-Ruiz, M. E., & Richter, R. L. (1997). Effect of homogenization pressure on themilk fat globule membrane proteins. Journal of Dairy Science, 80, 2732e2739.

Canselier, J. P., Delmas, H., Wilhelm, A. M., Abismail, B., Canselier, H., Delmas, A. M.,et al. (2002). Ultrasound emulsificationdan overview. Journal of DispersionScience and Technology, 23, 333e349.

Carter, J. (1993). Flax seed as a source of alpha-linolenic acid. Journal of the AmericanCollege of Nutrition, 12, 551.

Chanamai, R., & McClements, D. J. (2000). Impact of weighting agents and sucroseon gravitational separation of beverage emulsions. Journal of Agricultural andFood Chemistry, 48, 5561e5565.

Chandrapala, J., Martin, G. J. O., Zisu, B., Kentish, S. E., & Ashokkumar, M. (2012). Theeffect of ultrasound on casein micelle integrity. Journal of Dairy Science, 95,6882e6890.

Chandrapala, J., Zisu, B., Palmer, M., Kentish, S., & Ashokkumar, M. (2011). Effects ofultrasound on the thermal and structural characteristics of proteins in recon-stituted whey protein concentrate. Ultrasonics Sonochemistry, 18, 951e957.

Corredig, M., & Dalgleish, D. G. (1998). Characterization of the interface of an oil-in-water emulsion stabilized by milk fat globule membrane material. Journal ofDairy Research, 65, 465e477.

Couedelo, L., Boue-Vaysse, C., Fonseca, L., Montesinos, E., Djoukitch, S., Combe, N.,et al. (2011). Lymphatic absorption of alpha-linolenic acid in rats fed flax seedoil-based emulsion. British Journal of Nutrition, 105, 1026e1035.

Dapcevic Hadnadev, T., Dokic, P., Krstonosic, V., & Hadnadev, M. (2013). Influ-ence of oil phase concentration on droplet size distribution and stability ofoil-in-water emulsions. European Journal of Lipid Science and Technology, 115,313e321.

Dickinson, E. (2001). Milk protein interfacial layers and the relationship to emulsionstability and rheology. Colloids and Surfaces B: Biointerfaces, 20, 197e210.

Garti, N., & Yuli-Amar, I. (2008). Micro- and nano-emulsions for delivery of func-tional food ingredients. In N. Garti (Ed.), Delivery and controlled release of bio-actives in foods and nutraceuticals (pp. 149e152). UK: Woodhead Publishing.

Gutierrez, G., Rayner, M., & Dejmek, P. (2009). Production of vegetable oil in milkemulsions using membrane emulsification. Desalination, 245, 631e638.

Harjinder, S. (2006). The milk fat globule membrane e a biophysical system for foodapplications. Current Opinion in Colloid and Interface Science, 11, 154e163.

Jafari, S. M., He, Y. H., & Bhandari, B. (2006). Nano-emulsion production by soni-cation and microfluidization e a comparison. International Journal of FoodProperties, 9, 475e485.

Jafari, S. M., He, Y., & Bhandari, B. (2007). Effectiveness of encapsulating bio-polymers to produce sub-micron emulsions by high energy emulsificationtechniques. Food Research International, 40, 862e873.

Joshi, K., Lad, S., Kale, M., Patwardhan, B., Mahadik, S. P., Patni, B., et al. (2006).Supplementation with flax oil and vitamin C improves the outcome of AttentionDeficit Hyperactivity Disorder (ADHD). Prostaglandins, Leukotrienes and Essen-tial Fatty Acids, 74, 17e21.

Karbstein, H., & Schubert, H. (1995). Developments in the continuous mechanicalproduction of oil-in-water macro-emulsions. Chemical Engineering and Pro-cessing, 34, 205e211.

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

Kuhn, K. R., & Cunha, R. L. (2012). Flax seed oilewhey protein isolate emulsions:effect of high pressure homogenization. Journal of Food Engineering, 111, 449e457.

Lee, J. (2005). The behaviour of ultrasound generated bubbles in the presence of surfaceactive solutes (PhD thesis). Australia: University of Melbourne.

Leong, T., Ashokkumar, M., & Kentish, S. (2011). The fundamentals of power ultra-sound: a review. Acoustics Australia, 39, 54e63.

Leong, T. S. H., Wooster, T. J., Kentish, S. E., & Ashokkumar, M. (2009). Minimising oildroplet sizeusingultrasonic emulsification.Ultrasonics Sonochemistry,16, 721e727.

Maali, A., & Mosavian, M. T. H. (2013). Preparation and application of nano-emulsions in the last decade (2000e2010). Journal of Dispersion Science andTechnology, 34, 92e105.

Madadlou, A., Mousavi, M. E., Emam-Djomeh, Z., Ehsani, M., & Sheehan, D. (2009).Comparison of pH-dependent sonodisruption of re-assembled casein micellesby 35 and 130 kHz ultrasounds. Journal of Food Engineering, 95, 505e509.

Mantzioris, E., James, M. J., Gibson, R. A., & Cleland, L. G. (1994). Dietary sub-stitution with an alpha-linolenic acid-rich vegetable oil increases eicosa-pentaenoic acid concentrations in tissues. American Journal of ClinicalNutrition, 59, 1304e1309.

Matsumiya, K., Takahashi, W., Inoue, T., & Matsumura, Y. (2010). Effects of bacte-riostatic emulsifiers on stability of milk-based emulsions. Journal of Food Engi-neering, 96, 185e191.

Mazza, M., Pomponi, M., Janiri, L., Bria, P., & Mazza, S. (2007). Omega-3 fatty acidsand antioxidants in neurological and psychiatric diseases: an overview. Progressin Neuro-Psychopharmacology & Biological Psychiatry, 31, 12e26.

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

Mirhosseini, H. (2010). Effect of various hydrocolloids on physicochemical charac-teristics of orange beverage emulsion. Journal of Food, Agriculture & Environ-ment, 8, 308.

Mohammadzadeh, H., Koocheki, A., Kadkhodaee, R., & Razavi, S. M. A. (2013).Physical and flow properties of d-limonene-in-water emulsions stabilized withwhey protein concentrate and wild sage (Salvia macrosiphon) seed gum. FoodResearch International, 53, 312e318.

Mora-Huertas, C. E., Fessi, H., & Elaissari, A. (2010). Polymer-based nanocapsules fordrug delivery. International Journal of Pharmaceutics, 385, 113e142.

Nakai, S. (1983). Structureefunction relationships of food proteins with anemphasis on the importance of protein hydrophobicity. Journal of Agriculturaland Food Chemistry, 31, 676e683.

A. Shanmugam, M. Ashokkumar / Food Hydrocolloids 39 (2014) 151e162162

Nakai, S., Cheung, E., & Voutsinas, L. P. (1983). Relationships of hydrophobicity toemulsifying properties of heat denatured proteins. Journal of Food Science, 48, 26.

Pearce, K. N., & Kinsella, J. E. (1978). Emulsifying properties of proteins: evaluationof a turbidimetric technique. Journal of Agricultural and Food Chemistry, 26, 716e723.

Roesch, R. R., Rincon, A., & Corredig, M. (2004). Emulsifying properties of fractionsprepared from commercial buttermilk by microfiltration. Journal of Dairy Sci-ence, 87, 4080e4087.

Romero, A., Beaumal, V., David-Briand, E., Cordobes, F., Guerrero, A., & Anton, M.(2011). Interfacial and oil/water emulsions characterization of potato proteinisolates. Journal of Agricultural and Food Chemistry, 59, 9466e9474.

Santana, R. C., Perrechil, F. A., & Cunha, R. L. (2013). High- and low-energy emul-sifications for food applications: a focus on process parameters. Food Engi-neering Reviews, 5, 107e122.

Sarkar, A., Horne, D. S., & Singh, H. (2010). Interactions of milk protein-stabilized oil-in-water emulsions with bile salts in a simulated upper intestinal model. FoodHydrocolloids, 24, 142e151.

Schubert, H., & Engel, R. (2004). Product and formulation engineering of emulsions.Chemical Engineering Research and Design, 82, 1137e1143.

Shanmugam, A., Chandrapala, J., & Ashokkumar, M. (2012). The effect of ultrasoundon the physical and functional properties of skim milk. Innovative Food Science& Emerging Technologies, 16, 251e258.

Soria, A. C., & Villamiel, M. (2010). Effect of ultrasound on the technological prop-erties and bioactivity of food: a review. Trends in Food Science & Technology, 21,323e331.

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, 7444e7458.

Tcholakova, S., Denkov, N. D., Ivanov, I. B., & Campbell, B. (2002). Coalescence inb-lactoglobulin-stabilized emulsions: effects of protein adsorption and dropsize. Langmuir, 18, 8960e8971.

Vankova, N., Tcholakova, S., Denkov, N. D., Ivanov, I. B., Vulchev, V. D., & Danner, T.(2007). Emulsification in turbulent flow: 1. Mean and maximum drop diametersin inertial and viscous regimes. Journal of Colloid And Interface Science, 312,363e380.

Wade, T., & Beattie, J. K. (1997). Electroacoustic determination of size and zetapotential of fat globules in milk and cream emulsions. Colloids and Surfaces B:Biointerfaces, 10, 73e85.

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

Wulff-Perez, M., Torcello-Gomez, A., Galvez-Ruiz, M. J., & Martin-Rodriguez, A.(2009). Stability of emulsions for parenteral feeding: preparation and charac-terization of o/w nanoemulsions with natural oils and Pluronic f68 as surfac-tant. Food Hydrocolloids, 23, 1096e1102.