IMPROVEMENT OF OXIDATIVE STABILITY OF RICE BRAN OIL EMULSION BY CONTROLLING DROPLET SIZE

IMPROVEMENT OF OXIDATIVE STABILITY OF RICE BRAN OILEMULSION BY CONTROLLING DROPLET SIZEjfpp_633 1..13

HOANG HAI NGUYEN1, KYEONG-OK CHOI1, DONG EUN KIM1, WIE-SOO KANG2 and SANGHOON KO1,3

1Department of Food Science and Technology, Sejong University, 98 Gunja-dong, Gwangjin-gu, Seoul 143-747, Korea2Department of Plant Biotechnology, Kangwon National University, Chuncheon, Korea

3Corresponding author.TEL: +82-2-3408-3260;FAX: +82-2-3408-4319;EMAIL: [email protected]

Received for Publication February 15, 2011Accepted for Publication August 26, 2011

doi:10.1111/j.1745-4549.2011.00633.x

ABSTRACT

Well-planned emulsification studies are needed to improve oxidative stability of ricebran oil. Factors that affect the droplet size during emulsification of rice bran oilinclude the amount of rice bran oil content, types of emulsifiers, power intensity andduration of sonication. In this study, the emulsions were prepared at different oilconcentrations (10, 15 and 20%), using different emulsifiers (Tween 80 and its mix-tures with Span 80), and subsequently sonicated at different durations (5–50 min).Longer sonication time resulted in smaller-sized emulsion droplets. The emulsionswith the mixture of Tween 80 and Span 80 formed smaller droplets and better oxida-tive stability than those with Tween 80 only. In conclusion, rice bran oil emulsionwith smaller droplets improves oxidative stability. This study may be helpful for tai-loring oxidative-stable emulsions with well-defined droplet size distribution.

PRACTICAL APPLICATIONS

Recently, food manufacturers have been interested in utilization of the polishingby-products such as rice bran. Rice bran oil contains well-balanced saturated andunsaturated fats and provides a good source of vitamin E, antioxidants, oryzanol andother micronutrients. However, food application of rice bran oil is limited because oflipid oxidation. One of the promising technologies is nanoemulsion fabrication,which is being applied to enhance the oral bioavailability and oxidative stability ofthe poorly water-soluble bran oils. Herein, attention is focused on improving oxida-tive stability by controlling droplet size through adding mixture of surfactants in theoil-in-water emulsion. Narrow size distribution of droplets is a challenge forimproved oxidative stability. In this study, proper formulations such as a mixture ofemulsifiers provide a resolution to overcome the oxidative stability limitations inutilization of rice bran oils.

INTRODUCTION

Consumers have recently become aware that plant oilsincluding rice bran oil have important health benefits, andnew research in this field is attracting immense public inter-est. Rice bran oil contains well-balanced saturated andunsaturated oils and provides an excellent source of vitaminE, antioxidants, oryzanol and other micronutrients(Monsoor et al. 2003). Rice bran oil has a number of poten-tial uses as functional emulsions in many fields such as thecosmetic, pharmaceutical and food industries (Hasenhuettland Hartel 2008). One of the promising technologies is

nanoemulsion fabrication, which is being applied to enhancethe oral bioavailability and gravitic stability of poorly solublecomponents. Nanoemulsions because of small droplet sizehave a higher solubilization capacity than simple micellarsolutions.

Food applications of rice bran oil can be expanded when itshealth benefits are improved by proper processing tech-niques. Droplet size usually relates to functionality of foodemulsions (Raikar et al. 2009). An emulsion with smaller-sizedroplets has a larger surface area and is desirable for processessuch as improved water absorption, flavor release and bio-availability (Aungst 1993). However, food application of rice

Journal of Food Processing and Preservation ISSN 1745-4549

1Journal of Food Processing and Preservation •• (2012) ••–•• © 2012 Wiley Periodicals, Inc.

bran oil is limited because of lipid oxidation during storage;lipid oxidation is a major cause of quality deterioration in ricebran oil products by altering odor and nutritional quality.The droplets’ characteristics may affect the oxidation kineticsin oil-in-water (o/w) emulsion, depending on their size andphysical state (Osborn and Akoh 2004). Lipid oxidation isaccelerated by reactions that take place at the surface of emul-sion droplets; smaller droplets have an increased specificsurface area and are more prone to oxidation than largerdroplets (Lethuaut et al. 2002). Nanoemulsion, which hasincreased specific surface area, may not be considered a goodsystem to retard lipid oxidation. However, surface character-istics of the droplets are more important than their size.Several authors have found that surface charge can influencethe oxidative stability of lipid droplets in emulsion systems(Yoshida and Niki 1992; Mei et al. 1998). Interfacial proper-ties of the lipid droplets could be the important determinantsof the degree of oxidation in an emulsion. The physical char-acteristics of the droplets may also affect the oxidative stabil-ity of o/w emulsion depending on their concentration, size,physical state and charge (Mancuso et al. 1999). However,determination of the optimal emulsion preparation condi-tion has been difficult. Regression analysis allows for anapproximate fit by minimizing the gap between the datapoints and the fitted curve. Thus, the regression analysis ofexperimental data may help determine the process conditionsto obtain an emulsion with the desired physicochemical char-acteristics. Herein, we hypothesized that a smaller-sizedroplet with proper surface characteristics may be a favorablestructure to enhance gravitic aqueous stability and retard oxi-dation of rice bran oil.

Uniform and narrow size distribution of droplets isrequired for better control of functionality and productquality. Emulsion droplet size depends on the nature of theemulsifiers, the oil and aqueous phase, temperature,mechanical energy (high shear, homogenizer, centrifugation,ultrasound generators) and emulsification conditions (Dal-gleish et al. 1995; Dickinson 2009). Emulsifiers also play a rolein the oxidative stability of oil droplets (Fomuso et al. 2002).The primary role of the emulsifier is to adsorb at the surface ofthe freshly formed droplets and concomitantly prevent themfrom coalescing with their neighboring droplets (Lee et al.2009). Tween 80 (hydrophilic-lipophilic balance [HLB] 15)and Span 80 (HLB 4.3) are the most commonly used emulsi-fiers that improve the stability of emulsions by forming a pro-tective layer around the droplets. Tween 80 is also used as anemulsifier in foods (Goff 2000). Tween 80 and Span 80 arenonionic surface active agents, which are generally consid-ered to be safe. Emulsifiers control interdroplet forces,thereby either preventing or retarding the rate of coalescenceof colloidal droplets during emulsion formation (Wulff-Pérez et al. 2009). Tween 80 is soluble in water and is thereforeused to form o/w emulsions, while Span 80 can be used to

form o/w emulsions. Combinations of these emulsifiers mayaid in the formation of well-stabilized emulsions. Because ofthe closely related chemical characteristics of Tween 80 andSpan 80, they complement each other in terms of both hydro-philic and hydrophobic interactions. When used together, themixture of Tween 80 and Span 80 can pack more denselyalong the interface between the oil and aqueous phase (Ingridet al. 2007). Because it is usually desired to obtain emulsionshaving maximum oxidative stability and well-defined dropletcharacteristics, proper formulations such as a mixture ofTween 80 and Span 80 may have the potential for approachingoptimal conditions during emulsification.

Herein, attention is focused on improving oxidative stabil-ity by controlling droplet size simply through adding amixture of surfactants in the o/w emulsion. In addition,droplet size versus sonication time is modeled to predict theoptimal emulsion preparation conditions for obtaining con-trolled droplet characteristics and oxidative stability in branoil emulsions.

MATERIALS AND METHODS

Bran Oil Emulsion Preparation

An emulsion was prepared by adding slowly distilled waterto a mixture of rice bran oil (Cosmetic Aromanara Co.,Seoul, Korea) and emulsifier under a magnetic stirrer. HCl(0.1 N, Deajung Chemical & Metals Co., Shiheung, Korea)was added to adjust the pH of the emulsion to 7.0 in a phos-phate (5 mM; pH 7.0) buffer solution. The temperature waskept at 25C. The varying parameters of the emulsion prepa-rations were bran oil contents, type of emulsifiers, and soni-cation durations. Different concentrations of bran oil at 10,15 and 20% wt were studied. For o/w emulsion systempreparation, three combinations of the emulsifiers’ compo-sitions containing 6% wt Tween 80 (polyoxyethylene sorbi-tan monooleate, HLB 15, Tokyo Chemical Industry Co.,Tokyo, Japan), mixture of 4% wt Tween 80 and 2% wt Span80 (sorbitan monooleate, HLB 4.3, Tokyo Chemical Indus-try Co.), and mixture of 3% wt Tween 80 and 3% wt Span80 were investigated.

Ultrasonication experiments employed a Sonics VCX 750(Sonics & Materials Inc., Newtown, CT) with a 6-mm diam-eter tip, which was placed in a glass tube with an internaldiameter of 50 mm with an in-built cooling jacket. An ultra-sound having a power of 562 W was applied for 5, 15, 30, 40and 50 min, and corresponding accumulative powerswere 46.8, 140.5, 281.0, 374.6 and 468.3 Wh, respectively.Chilled water was passed continuously through the jacket toavoid temperature increase during the sonication process.The levels of independent parameters are presented inTable 1.

RICE BRAN OIL EMULSION H.H. NGUYEN ET AL.

2 Journal of Food Processing and Preservation •• (2012) ••–•• © 2012 Wiley Periodicals, Inc.

Size Distribution and Zeta-Potentialof Droplets

Droplet size distribution and zeta-potential were determinedto observe the emulsion stability using particle size analyzer(Delsa nano C, Beckman Coulter, Inc., Brea, CA). Sampleswere diluted with distilled water according to instructionsgiven in the service manual and 3 mL of diluted samples wereplaced in a glass cell. All samples were measured three times atfixed temperature and angles of 25 and 165°, respectively. Theparticle size analyzer provided mean particle size distribu-tion. The samples mentioned earlier were further diluted andadjusted to pH 4.0 in order to measure the zeta-potential. Thesample was injected into a flow cell in which its zeta-potentialwas measured.

Nonlinear Regression

Regression analysis of the experimental data was performedbased on nonlinear models. A dependent variable (dropletsize) and one or more independent variables (sonicationtime) were set as the numerical data. The droplet size in theregression equation was modeled as a function of the sonica-tion time. The parameters were estimated to give the best fit ofthe data. The Sigma Plot (Ver. 10.0, Systat Software Inc., SanJose, CA) was used for nonlinear regression analysis. The datawas analyzed to determine the best predictive model betweenthe emulsion preparation condition and the droplet size. Inaddition, Duncan’s multiple tests were used to compare theminimum droplet sizes predicted by the nonlinear models.

Oxidative Stability

Lipid hydroperoxides were measured to evaluate oxidative sta-bility of the rice bran oil emulsions (Shantha and Decker1994).In brief,all of the samples were stored at 50°C to acceler-ate the lipid oxidation process. Lipid hydroperoxides weremeasured by mixing 3.4 mL of an emulsion sample with25 mLof isooctane/2-propanol(3:1,v/v)mixturebyvortexingfor 30 s and subsequently isolating the organic solvent phaseby centrifugation at 10,000 rpm for 10 min. The organicsolvent phase (200 mL) was added to 2.8 mL of methanol/butanol (2:1, v/v) mixture, followed by 15 mL of 3.97 M

ammonium thiocynate and 15 mL of ferrous iron solution,which were prepared by mixing 0.132 M BaCl2 and 0.144 MFeSO4. Then, the solution was placed into a test tube andallowed to oxidize at 37C in the dark for 30 min. The absor-bance of the solution was measured at 510 nm (DU 730, Lifescience UV/Vis Spectrophotometer, Beckman Coulter, Inc.,Brea, CA). Hydroperoxide concentrations were determinedusing a standard curve made from cumene hydroperoxide.

Statistical Analysis

Data in triplicate were analyzed by one-way analysis of vari-ance using a SAS program (SAS Institute Inc., Cary, NC). Sta-tistical significance was considered at P < 0.05.

RESULTS AND DISCUSSION

Effect of Emulsifiers on Droplet Size in BranOil Emulsion

The dynamic light scattering particle size analyzer measuredthe droplet size of the bran oil emulsion prepared as shown inFig. 1.The size distribution of the droplets in the emulsion wasobtained and corresponding mean size was calculated fromthe particle size analyzer. The average droplet size of the branoil emulsion depended on the types and the compositions ofthe emulsifiers as shown in Fig. 2. The results indicate that themixturepreparedwithTween80andSpan80providedsmallerdroplet sizes than that emulsified with only Tween 80. When10% oil concentration was used, the droplet size of the emul-sion prepared with mixture of 4% Tween 80 and 2% Span 80was the smallest among the three emulsifier combinationstested at the early sonication stage, but the rate of droplet sizereduction decelerated after 20 min.When the sonication timeincreased, the droplet size decreased after increasing the ratioof Span80intheemulsifierused.Resultsusingthe15%oilcon-centration showed a similar pattern to that of the 20% oil.

As shown in the Fig. 2, increasing the oil concentrationfrom 10 to 20% led to an increase of the droplet size in theemulsion. The emulsions with 10% oil content and themixture of the emulsifiers were stabilized quickly over timeunder the ultrasonication treatment because proper combi-nation of hydrophilic and hydrophobic emulsifiers preventedseparation of the oil from water phases. The emulsion systemwith high oil concentration (20%) required an increasedquantity of hydrophobic emulsifier to disperse the oil com-pletely and for further stabilization. Moreover, the mixture ofTween 80 and Span 80 packed more densely along the inter-face between water and oil phases, and the emulsion stabiliza-tion capacity of the mixture was relatively stronger thanindividual emulsifiers tested. As shown in Fig. 1, it can also beobserved that emulsion with 3% Tween 80 and 3% Span 80has narrow droplet size distribution compared with other

TABLE 1. LEVELS OF INDEPENDENT PARAMETERS OF BRAN OILEMULSION PREPARATION

Independent variables Values of levels

Oil content (%) 10; 15; 20Ratio of Tween 80 to

Span 806:0 (HLB 15); 4:2 (HLB 11.4); 3:3 (HLB 9.6)

Time (min)/ultrasonicationpower (Wh)

5/46.8; 15/140.5; 30/281.0; 40/374.6;50/468.3

H.H. NGUYEN ET AL. RICE BRAN OIL EMULSION


emulsions. The droplet size in all samples decreased with theincreasing sonication time for the emulsions prepared at alloil concentrations tested.

The HLB values of the emulsifier mixtures (HLBmix) werecalculated using weight fraction of the corresponding emulsi-fiers as shown in the following:

HLB HLB HLBmix A A B B= × + ×f f ,

where HLBA and HLBB are the HLB values of components Aand B, respectively, and fA and fB are the weight fractions ofcomponents A and B, respectively (Wang et al. 2009). TheHLBmix values of Tween 80 only, mixture of 4% Tween 80 and2% Span 80, and mixture of 3% Tween 80 and 2% Span 80were 15, 11.4 and 9.6, respectively (see Table 1).

The effect of mixed emulsifiers has also been studied byother researchers. The effect of weight ratios of differenthydrophobic emulsifiers (Span 20, 40, 60 and 80) to a hydro-philic emulsifier (Tween 80) has been studied during micro-emulsion preparation (Cho et al. 2008). The mixtures ofemulsifiers provided better microemulsion storage stabilitythan Tween 80 alone. The stability of the emulsions has beenshown to improve by using a combination of emulsifiers,because the hydrophilic-lipophilic properties of an emulsifierare balanced by using a mixture (optimal HLB). The optimalHLB provides only a minimal interfacial tension between theoil and aqueous phase (Forster et al. 1992, 1994). Hence, at theoptimal HLB of the emulsion system, the smallest droplet canbe formed.

A

C

B

FIG. 1. DROPLET SIZE DISTRIBUTION OF THE BRAN OIL EMULSIONS PREPARED AT (A) 6% TWEEN 80, (B) 4% TWEEN 80 – 2% SPAN 80 AND (C) 3%TWEEN 80 – 3% SPAN 80



Zeta-Potential

The result of the zeta-potential measurement at pH 4 isshown in Fig. 3. The zeta-potentials of all samples displayeda negative value and were dependent on the surfactant type.The emulsion stabilized by the surfactant mixture of Tween80 and Span 80 indicated larger negative zeta-potentialvalue compared with either emulsion without surfactant orTween 80-added emulsion at all oil contents. In the disper-sion system, particles or droplets have charges on theirsurface because of the selective adsorption of ions, includingprotons (Lin and Chaudhury 2008). Zeta-potential is auseful indicator to understand the behavior and interactionof particles and droplets in an aqueous solution, which isaffected by factors such as pH, electrolytes and surfactant(Karraker and Radke 2002). In this study, the emulsionformed with Tween 80 only showed the zeta-potential value

close to zero, ranging -4 to -9 mV. The result resembledthose of an earlier study (Kato et al. 2009), in which it wasreported that the stability of a dispersion stabilized byTween 80 increases in spite of displaying a low zeta-potentialvalue because the surfactant adsorbed on the surface of dis-persion has a sterical interaction and small electrostaticinteraction between their molecules; the sterical repulsivepower of surfactant molecules could prevent the agglomera-tion between droplets or particles.

Nonlinear Regression of Droplet Size as aFunction of Sonication Time in BranOil Emulsion

From Fig. 4, it can be seen that the droplet size decreases asthe sonication time increases. In the method used here, the

Sonication time (min)

Dro

plet

siz

e (n

m)

0

200

400

600

800

6% Tween 804% Tween 80 - 2% Span 803% Tween 80 - 3% Span 80


0 10 20 30 40 50 0 10 20 30 40 50

0 10 20 30 40 50

Dro

plet

siz

e (n

m)

0

100

200

300

400

500

600

700

6%Tween 80 4%Tween 80 - 2% Span 80 3% Tween 80 - 3% Span 80


Dro

plet

siz

e (n

m)

0

100

200

300

400

500

600

700

6% Tween 804% Tweemn 80 - 2% Span 803% Tweemn 80 - 3% Span 80

A B

C

FIG. 2. EFFECT OF EMULSIFIERS ON DROPLET SIZE OF BRAN OIL EMULSIONS PREPARED AT (A) 10, (B) 15 AND (C) 20% OIL CONTENT



intensive ultrasound supplied the power needed to dispersesmall droplets of the bran oil phase into the water phase. Inthe dispersing zone, the imploding cavitations caused inten-sive shock waves through the bubbles in the surroundingliquid, resulting in the formation of liquid jets with highliquid velocity (Gohtani et al. 1999).

Several mathematical models were used to fit the curve ofthe emulsion droplet size versus sonication time. Because thedroplet size decreased nonlinearly with the sonication timeduring emulsification, three mathematical models werechosen to present the decay shape (Jena and Das 2006; Ko andGunasekaran 2008):

Model I y = ae-bx

Model II y = y0 + ae-bx

Model III y = aex+c

Where parameters x (the sonication time) and y (thedroplet size) are independent and dependent variables,respectively; a, b, c and y0 are coefficients.

The goodness-of-fit of the chosen nonlinear models wastestedbyR2,whichshowedvalues from0.0to1.0.AnR2 valueof0.0 means that knowing x is not useful to predict y. There wasno linear relationship between x and y. When R2 equals 1.0, allpoints lie exactly on a straight line with no scatter, and thus

Surfactant

NS 3% Tween 80-3% Span 80

Zet

a-po

tent

ial (

mV

)

-20

-18

-16

-14

-12

-10

-8

-6

-4

-2

0

a

b

bc c

Surfactant

Zet

a-po

tent

ial (

mV

)

-20

-18

-16

-14

-12

-10

-8

-6

-4

-2

0

a

b

c c

-20

-18

-16

-14

-12

-10

-8

-6

-4

-2

0

c

a

b

c

A B

C

4% Tween 80-2% Span 80

6% Tween 80



6% Tween 80



6% Tween 80

FIG. 3. ZETA-POTENTIALS AT PH 4 OF BRAN OIL EMULSIONS PREPARED AT (A) 10, (B) 15, AND (C) 20% OIL CONTENTLetters a, b, and c found in Figure 3 are duncan’s groupings at the same oil content.



knowingxpredictsy.Thesumof squares(SS)andmeansquare(MS) of the models were compared, and results are listed inTable 2.We observed that R2 of Model II is larger than the R2 ofModel I but similar to that of Model III. Because the R2 ofModel I was not acceptable for the curve fitting, Model I wasrejected. At infinite x (long-time sonication), the nonlinearregression from Model III has a horizontal asymptote y = a,which means that coefficient a is the minimum droplet size.Minimum droplet sizes calculated from Models II (coefficienty0) and III (coefficient a) were compared with those of theexperimental data, and the results are listed in Table 3. Dun-can’s multiple tests were used to compare the minimumdroplet sizes. The minimum droplet size y0 of Model II is com-parable with the results of the experimental data. Simulta-neously, compared with Model III, SS and MS values of ModelII were low. In conclusion, Model II was chosen to fit the non-linear curve of droplet size as a function of sonication time.

The coefficients were calculated using Model II(y = y0 + ae-bx) as shown in Fig. 5. The mathematical mean-

ings for coefficients y0, a and b of Model II can be easily inter-preted. For Model II, at a larger x value (i.e., infinite x),e-bx = 0, at which the curve has a horizontal asymptote y = y0.Thus, with a larger x value, y0 is the minimum droplet size.The initial droplet size before sonication can be representedusing coefficients y0 and a. The initial droplet size is y0 + a,because y = y0 + a when x = 0. In general, for fine emulsionpreparations, the minimum droplet size is small (i.e., y0 << a,y0 + a ª a), so the initial droplet size can be defined as coeffi-cient a. In addition, coefficient b of Model II means the stiff-ness of the curve.

The effect of the types and the compositions of the emulsi-fiers at different bran oil contents can be easily explained byusing the coefficients y0, a, and b of Model II. Coefficient a(initial droplet size) and coefficient y0 (minimum dropletsize) were the smallest when the Span 80 concentration wasincreased to 3%. This model fits well for the emulsion prepa-ration conditions (10, 15 and 20% oil content) and the com-binations of emulsifier compositions. Based on the above


0 10 20 30 40 50 60 0 10 20 30 40 50 60

0 10 20 30 40 50 60

Dro

plet

siz

e (n

m)

200

300

400

500

600

700

800

10% wt of bran oil15% wt of bran oil20% wt of bran oil


Dro

plet

siz

e (n

m)

100

200

300

400

500

600

700

10% wt of bran oil15% wt of bran oil20% wt of bran oil


Dro

plet

siz

e (n

m)

100

200

300

400

500

600

700

10% wt of bran oil15% wt of bran oil20 % wt of bran oil

A B

C

FIG. 4. EFFECT OF THE SONICATION TIME ON THE DROPLET SIZE OF THE BRAN OIL EMULSIONS PREPARED AT (A) 6% TWEEN 80, (B) 4% TWEEN 80 –2% SPAN 80 AND (C) 3% TWEEN 80 – 3% SPAN 80



results, y = y0 + ae-bx was the best model to fit the nonlinearcurve of droplet size in terms of sonication time.

Effect of Sonication Time on Droplet Size inBran Oil Emulsion

The effect of sonication time on droplet size was prominent.From the initial stage of emulsification, droplet size decreasedsteeply as shown in Fig. 4. Even after 30 min, droplet size con-tinued to decline but at a slower rate.At the last emulsificationstage, the effect of sonication was not as strong as in the initialstage. Droplet size did not decrease significantly when sonica-tion time was over 40 min.

As the ratio of Span 80 to Tween 80 increased, the finaldroplet size of the emulsion decreased. A similar trend wasobserved with all emulsifier ratios in the emulsions preparedwith all tested oil concentrations (10, 15 and 20%). The emul-sion with 15% oil content especially provided the best corre-lation of the model and the experimental data in terms ofsonication time and droplet size. On the other hand, when theoil content was 20%, the correlation of the model and the

experimental data was not as good as that of 10 or 15%. In thecase of 20% oil content using Tween 80 and Span 80, themixture could not disperse completely in the aqueous phasebecause, in general, a larger amount of hydrophilic emulsifierwas needed for stabilization with a larger oil content (Dickin-son 2009).

Coefficient b was calculated using Model II means the stiff-ness of the curve of the emulsion droplet size versus sonica-tion time. The coefficient b is shown in Fig. 5. In other words,the stiffness of the plots of droplet size versus sonication timedepended significantly on the type of emulsifier and the ratioof emulsifier composition. The coefficient a of the model hasthe largest value for the emulsion prepared with the mixtureof 4% Tween 80 and 2% Span 80, but smallest value obtainedwith the mixture of 3% Tween 80 and 3% Span 80. Theminimum droplet size y0 of the model for the emulsion withTween 80 only was the largest; the value decreased consider-ably with the increased Span 80 content. The trends of thecoefficient y0 values of the emulsions prepared with 10, 15 and20% oil content were similar. These results are similar to thoseof a previous study (Jafari et al. 2006) in which increased

TABLE 2. THE FITNESS (R2, SS, MS) OF THE MODELS I, II, AND III AT DIFFERENT OIL CONTENTS AND TYPES OF EMULSIFIERS

Model Fitness

Oil content (%)

10 15 20

T6a T4-S2b T3-S3c T6 T4-S2 T3-S3 T6 T4-S2 T3-S3

I R2 0.96 � 0.02 0.94 � 0.01 0.98 � 0.01 0.92 � 0.00 0.90 � 0.01 0.96 � 0.02 0.61 � 0.02 0.95 � 0.02 0.87 � 0.03SS 10,712.8 6,681.7 27,509.2 19,050.6 22,000.1 3,031.9 12,161.9 5,684.2 6,620.9MS 824.1 513.9 1,528.3 1,465.4 1,692.3 233.2 935.4 437.2 509.3

II R2 1.00 � 0.00 0.99 � 0.00 0.98 � 0.07 1.00 � 0.00 1.00 � 0.00 1.00 � 0.00 1.00 � 0.00 0.96 � 0.02 0.88 � 0.03SS 303.15 683.74 1,368.87 246.6 968.4 373.6 9,133.4 817.5 722.6MS 151.6 56.9 693.4 20.5 80.7 31.1 761.1 69.8 80.2

III R2 1.00 � 0.00 0.99 � 0.01 0.98 � 0.01 1.00 � 0.00 1.00 � 0.00 1.00 � 0.00 1.00 � 0.00 0.96 � 0.02 0.87 � 0.03SS 1,665.7 1,102.4 N/Ad 618.4 718.8 388.1 9,820.7 1,267.9 1,120.9MS 138.8 91.8 N/A 51.3 59.9 32.3 818.4 105.6 93.4

a T6: 6% Tween 80 only.b T4-S2: 4% Tween 80 and 2% Span 80.c T3-S3: 3% Tween 80 and 3% Span 80.d N/A: the model was not converged but exceeded maximum number of iteration.MS = mean of squares; SS = sum of squares.

TABLE 3. MINIMUM DROPLET SIZE OF THE BRAN OIL EMULSION PREDICTED BY THE MODELS II AND III AT DIFFERENT TYPES OF THE EMULSIFIERS,AND 10 15 AND 20% OIL CONTENT

Minimum droplet size(nm)

Oil content (%)

10 15 20

T6A T4-S2B T3-S3C T6 T4-S2 T3-S3 T6 T4-S2 T3-S3

Experimentally measured 250A 172A 141A 336A 262A 207A 170A 260A 329A

Model II 219A 179A 64B 321A 253A 175A 170A 200B 304A

Model III 115B 89B N/A 228B 184B 99B 168A 84C 216B

A, B and C are Duncan’s groupings of the minimum droplet sizes at the same oil content and emulsifier.a N/A: the model was not converged but exceeded maximum number of iterations.



sonication time decreased droplet size during emulsificationof an o/w emulsion of d-limonene. Reducing the amountof dispersed phase (d-limonene) decreased the size ofemulsions.

Our predictive model showed a good fit for the bran oilemulsions prepared with different oil concentrations, thetypes of emulsifiers and the ratio of emulsifiers. The predic-tive model is also useful to determine required sonicationtime for formulation of emulsions with defined droplet size.

Effect of Engineering Parameters onOxidative Stability

Oxidative stability determined by hydroperoxide concentra-tion is shown in Fig. 6. Hydroperoxide concentrationincreased with increasing oil content in the emulsions. It isalso interesting to note that there was no significant increasein the hydroperoxide levels until 20 days, and thereafter a

drastic increase was observed in all samples. Emulsions withthe emulsifiers such as Tween 80 or mixture of emulsifierswere more stable from lipid oxidation than non-emulsifier-added emulsions. The emulsifiers surrounding oil dropletsact as a radical scavenger and a protective barrier that pre-vents oxidation-inducers from penetration and diffusioninside the droplets (Porter et al. 1995).

In the case of emulsions without an emulsifier, the effect ofhigh oil content on hydroperoxide increase was conspicuouson long-time storage. After 30 days of storage, the hydroper-oxide values of the emulsions with 10 and 15% oil contentincreased, but the difference compared with the emulsionwithout emulsifier was not large. However, the difference wasrelatively large in the emulsions with 20% oil content.

As mentioned previously, droplet size of the emulsion withonly Tween 80 was larger than that of the mixture of Tween 80and Span 80. As expected, the hydroperoxide values were lowin the emulsions prepared with 6% Tween 80. The emulsion

Oil content (%)

10 15 20

b va

lue

0.00

0.02

0.04

0.06

0.08

0.10


a

ab

a

b

c

b

a

a

Oil content (%)

025101

a va

lue

0

100

200

300

400

500

600

700

6% Tween 804% Tween 80 - 2% Span3% Tween 80 - 2% Span

a

b

c

a a

b a

b

a

Oil content (%)

10 15 20

Y0

valu

e

0

100

200

300


a b

c

a

b

c

a aa

A B

C

FIG. 5. COEFFICIENTS A, B AND Y0 OF THE MODEL II (Y = Y0 + AE-BX) FOR THE BRAN OIL EMULSION AT DIFFERENT OIL CONTENTS AND EMULSIFIERS’COMPOSITIONS



with Tween 80 (only hydrophilic emulsifier) could obtaingood oxidative stability because a higher concentration ofTween 80 would form a thicker interfacial membrane on thesurface of the droplet and the packing of the surfactant mol-ecules at interface would be tightened (Tadros et al. 2004);hence, this membrane acts as an efficient barrier to the diffu-sion of oxidative initiators into the oil droplets (McClementsand Decker 2000). However, beyond a 20-day storage period,hydroperoxide values of emulsion prepared with 6% Tween80 increased steeply. Different types and concentrations ofemulsifiers give different droplet sizes (Ko and Gunasekaran2006) that might also affect the oxidation stability.

The oxidative stability of the emulsions prepared with themixture of Tween 80 and Span 80 (mixture of hydrophilic andhydrophobic emulsifiers), may be better or worse than thoseprepared with Tween 80 only. In this study, the smallest drop-lets formed at 10% oil content and mixture of 4% Tween 80and 2% Span 80 showed higher hydroperoxide levels. On the

contrary, small droplets were prepared with the mixture ofTween 80 and Span 80, but showed low hydroperoxide levels(Fig. 7). A few studies have reported that decrease in dropletsize increases specific surface area, which is more prone tooxidation (Gohtani et al. 1999; Lethuaut et al. 2002); themixed emulsifiers decrease droplet size, but can increase lipidoxidation. However, the mixture of emulsifiers can provide apositive effect on the prevention of lipid oxidation. Additionof the hydrophobic component in the mixed emulsifierswould contribute to prevention of lipid oxidation (Abismaïlet al. 1999). Span 80 can prevent movement of water-solubleoxidant such as lipoxygenase. Herein, the droplet sizedecreased by increasing the ratio of Span 80 concentration inthe mixture of emulsifiers used, but there was no significantdifference in oxidative stability during short-term storage.Moreover, the relatively small droplets prepared at 3% Tween80 and 3% Span 80 mixture condition showed the best anti-oxidant efficiency (Fig. 7).

Time (day)

Hyd

rope

roxi

de c

once

ntra

tion

(ppp

m)

0

10

20

30

40

50

Control6% Tween 804% Tween 80 - 2% Span 803% Tween 80 - 3% Span 80

Time (day)

0day 5day 10day 20day 30day

0day 5day 10day 20day 30day 0day 5day 10day 20day 30day

Hyd

rope

roxi

de c

once

ntra

tion

(ppm

)

0

10

20

30

40

50Control6% Tween 804% Tween 80 - 2% Span 803% Tween 80 - 3% Span 80

Time (day)

Hyd

rope

roxi

de c

once

ntra

tion

(ppm

)

0

10

20

30

40

50Control6% Tween 804% Tween 80 - 2% Span 803% Tween 80 - 3% Span 80

BA

C

FIG. 6. OXIDATIVE STABILITY OF THE BRAN OIL EMULSIONS WITH 6% TWEEN 80, 4% TWEEN 80 – 2% SPAN 80 AND 3% TWEEN 80 – 3% SPAN 80AT (A) 10, (B) 15 AND (C) 20% OIL CONTENTS FOR 30-DAY STORAGE



There are several reports that the effect of droplet size onlipid oxidation is not uniform; some studies indicated thatsmaller droplet sizes led to higher oxidation rates because ofincreased surface area (Lee et al. 2009), but another reportedno dependence of lipid oxidation rate on droplet size(Roozen et al. 1994). Because limited amounts of hydroper-oxides were available in the systems, they might have allbeen present at the droplet surface in every o/w emulsionsystem studied.

Effect of emulsifier type on lipid oxidation is relatively sig-nificant. In this study, we used the nonionic emulsifiers withthe HLB values 15, 11.4, and 9.6, respectively. It was observedthat hydroperoxide value decreased with increasing HLBvalue of emulsifiers. Therefore, the effect of emulsifier com-position was more pronounced than droplet size on oxidationstability of bran o/w emulsion. The use of emulsifier mixturenot only improves gravitic aqueous stability through dropletsize reduction, but also controlled oxidation stability of theemulsion system. The result was in close agreement with pre-vious studies (Kubouchi et al. 2002).]. Generally, at higheremulsifier concentrations, the packing of emulsifier mol-ecules at the oil–water interface is tighter; hence, the mem-brane acts as an efficient barrier to the diffusion of lipidoxidation initiators into the oil droplets (Coupland andMcClements 1996). In another report, lower emulsifier levels(0.25%) showed significantly higher oxidation levels thanemulsions prepared with 1% emulsifier (P < 0.05) (Fomusoet al. 2002). It may be deduced that emulsifier concentrationrather than droplet size distribution caused changes in oxida-tion properties.

CONCLUSION

This study investigates the effects of emulsifier type and ratio,oil content, and sonication time on droplet size and oxidativestability. The mathematical model y = y0 + ae-bx is proposedfor the bran o/w emulsion based on the experimental obser-vation. The effect of emulsion preparation conditions ondroplet size distribution can be used to predict the sonicationtime required for a bran o/w emulsion with different ratios ofemulsifiers. Under the same circumstances, the droplet sizecan be deduced because of the sonication time. This modelcan be expanded to predict formulations for preparing theemulsions with defined droplet sizes for various food applica-tions. In addition, oxidative stability of emulsion was mainlyaffected by oil content, and emulsifier conditions with a smallamount of oil addition induced lower hydroperoxide values.In general, small size is not good for oxidative stability, but itincreases bioavailability of the system. Therefore, we over-come the problem by controlling emulsifier conditions,which give a good choice for both oxidative stability and bio-availability properties. For future applications, the dropletsize distribution could be used as an evaluation criterion forthe quality and stability of emulsions prepared under differ-ent preparation conditions.

ACKNOWLEDGMENTS

This study was carried out with the support of “CooperativeResearch Program for Agricultural Science & TechnologyDevelopment (Project No.200806A01036035),”RDA, Repub-

Oil content

10% 15% 20%

Hyp

erox

ide

conc

entr

atio

n (p

ppm

)

0

10

20

30

40

50

60

Control6% Tween 804% Tween 80 - 2% Span 803% Tween 80 - 3% Span 80

bc

ab

c

a

a

a

a

a

ab

ab

b

a

FIG. 7. HYDROPEROXIDE CONCENTRATIONOF THE BRAN OIL EMULSIONS WITH 6%TWEEN 80, 4% TWEEN 80–2% SPAN 80 AND3% TWEEN 80–3% SPAN 80 AT 10, 15 AND20% OIL CONTENTS



lic of Korea and Korea Science and Engineering Foundation(KOSEF) grant funded by the Korean government (MEST)(grant code: F1706).

REFERENCES

ABISMAÏL, B., CANSELIER, J.P., WILHELM, A.M., DELMAS,H. and GOURDON, C. 1999. Emulsification by ultrasound:Drop size distribution and stability. Ultrason. Sonochem. 6,75–83.

AUNGST, B.J. 1993. Novel formulation strategies for improvingoral bioavailability of drugs with poor membrane permeationor presystemic metabolism. J. Pharm. Sci. 82, 979–987.

CHO, S., KIM, E.K., BAE, C.K. and MOK, J.P. 2008. Formulationof a cosurfactant-free o/w microemulsion using nonionicsurfactant mixtures. J. Food Sci. 73, E115–E121.

COUPLAND, J.N. and MCCLEMENTS, D.J. 1996. Lipidoxidation in food emulsions. Trends Food Sci. Technol. 7,83–91.

DALGLEISH, D.G., SRINIVASAN, M. and SINGH, H. 1995.Surface properties of oil-in-water emulsion droplets containingcasein and Tween 60. J. Agric. Food Chem. 43, 2351–2355.

DICKINSON, E. 2009. Hydrocolloids as emulsifiers and emulsionstabilizers. Food Hydrocolloids 23, 1473–1482.

FOMUSO, L.B., CORREDIG, M. and AKOH, C.C. 2002. Effect ofemulsifier on oxidation properties of fish oil-based structuredlipid emulsions. J. Agric. Food Chem. 50, 2957–2961.

FORSTER, T., SCHAMBIL, F. and VON RYBINSKI, W. 1992.Production of fine disperse and long-term stable oil-in-wateremulsions by the phase inversion temperature method. J.Dispersion Sci. Technol. 13, 183–193.

FORSTER, T., RYBINSKI, W.V., TESMANN, H. and WADLE, A.1994. Calculation of optimum emulsifier mixtures for phaseinversion emulsification. Int. J. Cosmet. Sci. 16, 84–92.

GOFF, H.D. 2000. Colloidal aspects of ice cream – A review. Int.Dairy J. 7, 363–373.

GOHTANI, S., SIRENDI, M., YAMAMOTO, N., KAJIKAWA, K.and YAMANO, Y. 1999. Effect of droplet size on oxidation ofdocosahexaenoic acid in emulsion system. J. Dispersion Sci.Technol. 20, 1319–1325.

HASENHUETTL, G.L. and HARTEL, R.W. 2008. Overview offood emulsifiers. In Food Emulsifiers and Their Applications(G.L. Hasenhuettl and R.W. Hartel eds.) pp. 1–9, Springer, NewYork, NY.

INGRID, A.M., MATT, G., ROB, V. and NICOLAAS, J.Z. 2007.Emulsions as delivery systems in foods. In Encapsulation andControlled Release Technologies in Food Systems (J.M. Lakkis,ed.) pp. 41–82, Blackwell Publising, Ames, IA.

JAFARI, S.M., HE, Y. and BHANDARI, B. 2006. Nano-emulsionproduction by sonication and microfluidization – Acomparison. Int. J. Food Prop. 9, 475–485.

JENA, S. and DAS, H. 2006. Modeling of particle size distributionof sonicated coconut milk emulsion: Effect of emulsifiers andsonication time. Food Res. Intern. 39, 606–611.

KARRAKER, K.A. and RADKE, C.J. 2002. Disjoining pressures,zeta potentials and surface tensions of aqueous non-ionicsurfactant/electrolyte solutions: Theory and comparison toexperiment. Adv. Colloid Interface Sci. 96, 231–264.

KATO, H., MIZUNO, K., SHIMADA, M., NAKAMURA, A.,TAKAHASHI, K., HATA, K. and KINUGASA, S. 2009.Observations of bound Tween80 surfactant molecules onsingle-walled carbon nanotubes in an aqueous solution.Carbon 47, 3434–3440.

KO, S. and GUNASEKARAN, S. 2006. Preparation of sub-100-nmb-lactoglobulin (BLG) nanoparticles. J. Microencapsul. 23,887–898.

KO, S. and GUNASEKARAN, S. 2008. Analysis of cheese meltprofile using inverse-Hill function. J. Food Eng. 87, 266–273.

KUBOUCHI, H., KAI, H., MIYASHITA, K. and MATSUDA, K.2002. Effects of emulsifiers on the oxidative stability of soybeanoil TAG in emulsions. J. Am. Oil Chem. Soc. 79, 567–570.

LEE, S-H., LEFÈVRE, T., SUBIRADE, M. and PAQUIN, P. 2009.Effects of ultra-high pressure homogenization on theproperties and structure of interfacial protein layer in wheyprotein-stabilized emulsion. Food Chem. 113, 191–195.

LETHUAUT, L., MÉTRO, F. and GENOT, C. 2002. Effect ofdroplet size on lipid oxidation rates of oil-in-water emulsionsstabilized by protein. J. Am. Oil Chem. Soc. 79, 425–430.

LIN, C.H. and CHAUDHURY, M.K. 2008. Usingelectrocapillarity to measure the zeta potential of a planarhydrophobic surface in contact with water and nonionicsurfactant solutions. Langmuir 24, 14276–14281.

MANCUSO, J.R., MCCLEMENTS, D.J. and DECKER, E.A. 1999.The effects of surfactant type, pH, and chelators on theoxidation of salmon oil-in-water emulsions. J. Agric. FoodChem. 47, 4112–4116.

MCCLEMENTS, D. and DECKER, E. 2000. Lipid oxidation inoil-in-water emulsions: Impact of molecular environment onchemical reactions in heterogeneous food systems. J. Food Sci.65, 1270–1282.

MEI, L., MCCLEMENTS, D.J., WU, J. and DECKER, E.A. 1998.Iron-catalyzed lipid oxidation in emulsion as affected bysurfactant, pH and NaCl. Food Chem. 61, 307–312.

MONSOOR, M., PROCTOR, A. and HOWARD, L. 2003.Aqueous extraction, composition, and functional properties ofrice bran emulsion. J. Am. Oil Chem. Soc. 80, 361–365.

OSBORN, H.T. and AKOH, C.C. 2004. Effect of emulsifier type,droplet size, and oil concentration on lipid oxidation instructured lipid-based oil-in-water emulsions. Food Chem. 84,451–456.

PORTER, N., CALDWELL, S. and MILLS, K. 1995. Mechanismsof free radical oxidation of unsaturated lipids. Lipids 30,277–290.

RAIKAR, N.B., BHATIA, S.R., MALONE, M.F., MCCLEMENTS,D.J., ALMEIDA-RIVERA, C., BONGERS, P. and HENSON,M.A. 2009. Prediction of emulsion drop size distributions withpopulation balance equation models of multiple dropbreakage. Colloids Surf. A Physicochem. Eng. Asp. 361,96–108.



ROOZEN, J.P., FRANKEL, E.N. and KINSELLA, J.E. 1994.Enzymic and autoxidation of lipids in low fat foods: Model oflinoleic acid in emulsified triolein and vegetable oils. FoodChem. 50, 39–43.

SHANTHA, N.C. and DECKER, E.A. 1994. Rapid, sensitive,iron-based spectrophotometric methods for determinationof peroxide values of food lipids. J. AOAC Int. 77,421–424.

TADROS, T., IZQUIERDO, P., ESQUENA, J. and SOLANS, C.2004. Formation and stability of nano-emulsions. Adv. ColloidInterface Sci. 108–109, 303–318.

WANG, L., DONG, J., CHEN, J., EASTOE, J. and LI, X. 2009.Design and optimization of a new self-nanoemulsifying drugdelivery system. J. Colloid Interface Sci. 330, 443–448.

WULFF-PÉREZ, M., TORCELLO-GÓMEZ, A., GÁLVEZ-RUÍZ,M.J. and MARTÍN-RODRÍGUEZ, A. 2009. Stability ofemulsions for parenteral feeding: Preparation andcharacterization of o/w nanoemulsions with natural oils andPluronic f68 as surfactant. Food Hydrocolloids 23, 1096–1102.

YOSHIDA, Y. and NIKI, E. 1992. Oxidation of methyl linoleate inaqueous dispersions induced by copper and iron. Arch.Biochem. Biophys. 295, 107–114.



Documents

IMPROVEMENT OF OXIDATIVE STABILITY OF RICE BRAN OIL EMULSION BY CONTROLLING DROPLET SIZE