Formation of nanoemulsions stabilized by model food-grade emulsifiers using high-pressure homogenization: Factors affecting particle size

Food Hydrocolloids 25 (2011) 1000e1008

Contents lists available at ScienceDirect

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

Formation of nanoemulsions stabilized by model food-grade emulsifiers using high-pressure homogenization: Factors affecting particle size Cheng Qian, David Julian McClements* Department of Food Science, University of Massachusetts Amherst, Amherst, MA 01003, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 April 2010 Accepted 24 September 2010

Nanoemulsions are finding increasing utilization in the food and beverage industries for certain applications because of their unique physicochemical and functional properties: high encapsulation efficiency; low turbidity; high bioavailability; high physical stability. In this study, we examined the impact of system composition and homogenization conditions on the formation of nanoemulsions using a highpressure homogenizer (microfluidizer). The mean particle diameter decreased with increasing homogenization pressure and number of passes, with a linear logelog relationship between mean particle diameter and homogenization pressure. The minimum droplet diameter that could be produced after 6 passes at 14 kbar depended strongly on emulsifier type and concentration: SDS < Tween 20 < b-lactoglobulin < sodium caseinate. Small-molecule surfactants formed smaller droplets than proteins, which was attributed to their ability to rapidly adsorb to the droplet surfaces during homogenization. The impact of phase viscosity was examined by using different octadecane-to-corn oil ratios in the oil phase and different glycerol-to-water ratios in the aqueous phase. The minimum droplet size achievable decreased as the ratio of disperse phase to continuous phase viscosities (hD/hC) decreased for SDSstabilized emulsions, but was relatively independent of hD/hC for b-lactoglobulin-stabilized emulsions. At low viscosity ratios, much smaller mean droplet diameters could be achieved for SDS (d w 60 nm) than for b-lactoglobulin (d w 150 nm). The information reported in this study will facilitate the rational design of food-grade nanoemulsions using high-pressure homogenization methods. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Nanoemulsion Microfluidization Emulsification conditions Viscosity ratio Emulsifier type

1. Introduction There is strong interest in the food and other industries in the use of nanoemulsions as delivery systems for non-polar functional components, such as lipophilic bioactive lipids, drugs, flavors, antioxidants, and antimicrobial agents (Hu, Johnston, & Williams, 2004; Kesisoglou, Panmai, & Wu, 2007; McClements, Decker, & Weiss, 2007; Sanguansri & Augustin, 2006; Weiss et al., 2008; Wissing, Kayser, & Muller, 2004). Oil-in-water nanoemulsions contain small oil droplets (d < 100 nm) dispersed within a watery continuous phase, with each oil droplet being surrounded by a protective coating of emulsifier molecules (Acosta, 2009; McClements et al., 2007; Tadros, Izquierdo, Esquena, & Solans, 2004). The stability, physicochemical properties and functional performance of nanoemulsion-based delivery systems can be controlled by altering their composition and preparation conditions so as to produce emulsions with different droplet concentrations,

* Corresponding author. E-mail address: [email protected] (D.J. McClements). 0268-005X/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodhyd.2010.09.017

compositions, particle size distributions and/or interfacial properties (Lesmes & McClements, 2009; McClements, 2010). A major potential advantage of nanoemulsions over conventional emulsions is that they can be made to be optically transparent by preparing droplets with dimensions much smaller than the wavelength of light (d < l) so that scattering is relatively weak (McClements, 2002a, 2010; Wooster, Golding, & Sanguansri, 2008). Consequently, they can be used to incorporate non-polar functional components into transparent aqueous-based food and beverage products. Nevertheless, this type of emulsion may become turbid or even opaque if droplet growth occurs during storage, e.g., due to flocculation, coalescence or Ostwald ripening. Hence, it is important to prevent these instability mechanisms from occurring in nanoemulsions after they have been formed. Nanoemulsions typically have much better stability to gravitational separation than conventional emulsions because the relatively small particle size means that Brownian motion effects dominate gravitational forces (McClements, 2005; Tadros et al., 2004). They also tend to have better stability against droplet flocculation and coalescence because the range of the attractive forces acting between the droplets decreases with decreasing particle size, while the range of

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008

the steric repulsion is less dependent on particle size (McClements, 2005; Tadros et al., 2004). Recent studies suggest that the bioavailability of encapsulated non-polar components is higher in nanoemulsions than conventional emulsions because of the small particle size and high surface-to-volume ratio (Acosta, 2009; Huang, Yu, & Ru, 2010). Hence, food-grade nanoemulsions may be particularly useful for increasing the bioactivity of lipophilic components that are normally poorly absorbed. A variety of preparation methods have been developed to prepare nanoemulsions, and these can conveniently be classified as either high energy or low energy approaches (Acosta, 2009; Leong et al., 2009; Tadros et al., 2004). High-energy approaches utilize mechanical devices capable of generating intense disruptive forces that breakup the oil and water phases and lead to the formation of tiny oil droplets, e.g., high-pressure valve homogenizers, microfluidizers and sonication methods (Gutierrez et al., 2008; Leong et al., 2009; Velikov & Pelan, 2008; Wooster et al., 2008). On the other hand, low energy approaches rely on the spontaneous formation of tiny oil droplets within mixed oilewater-emulsifier systems when the solution or environmental conditions are altered (Anton, Benoit, & Saulnier, 2008; Bouchemal, Briancon, Perrier, & Fessi, 2004; Chu, Ichikawa, Kanafusa, & Nakajima, 2007; Freitas, Merkle, & Gander, 2005; Tadros et al., 2004; Yin, Chu, Kobayashi, & Nakajima, 2008). The minimum size of the droplets that can be produced using each approach depends on many different factors. Previous studies have shown that the minimum particle size achievable using the highenergy approach depends on homogenizer type, homogenizer operating conditions (e.g., energy intensity, time, and temperature), sample composition (e.g., oil type, emulsifier type, and relative concentrations), and the physicochemical properties of the component phases (e.g., interfacial tension and viscosity). The overall objective of the current research is to establish the major factors that determine the size of the droplets produced using a high-energy approach: microfluidization. Fine emulsions can be produced using microfluidization by passing a coarse emulsion through an interaction chamber using a high-pressure pumping device (Jafari, He, & Bhandari, 2007). The interaction chamber consists of two flow channels, which are designed so that they cause two streams of the coarse emulsion to impinge on each other at high velocity, thus creating a very high shearing action that provides an exceptionally fine emulsion (Jafari, He, & Bhandari, 2006). Studies have shown that the particle size distribution produced by microfluidizers tends to be narrower and smaller that those produced by other homogenization devices (Jafari et al., 2006, 2007; PerrierCornet, Marie, & Gervais, 2005; Wooster et al., 2008). The final particle size achieved by a homogenizer is important because it determines the stability, appearance, texture, and bioavailability of the final product (Acosta, 2009; McClements et al., 2007). It is therefore important to establish the major factors that impact that particle size produced by different types of homogenizers. In this study, we systematically examined the influence of number of passes, homogenization pressure, emulsifier type, and disperse and continuous phase viscosities on the droplet size produced. Typically, the mean droplet diameter tends to decrease with increasing homogenization pressure and number of passes, but for certain emulsifiers (typically biopolymers) higher pressures and longer emulsification times may lead to “over-processing”, resulting in an increase in particle size (Jafari et al., 2007; Jafari, Assadpoor, He, & Bhandari 2008). Previous studies suggest that small-molecule surfactants adsorb to droplet surfaces more rapidly than biopolymers, which means they are more effective at producing the very small droplets in nanoemulsions (Azeem, Rizwan, Ahmad, Khar, et al., 2009a; Azeem, Rizwan, Ahmad, Iqbal, et al. 2009b; Jafari et al., 2007; Yuan, Gao, Zhao, & Mao, 2008). One of the aims of this paper was therefore to establish if

1001

biopolymers could be successfully used as emulsifiers to form foodgrade nanoemulsions. 2. Materials and methods 2.1. Materials Mazola Corn oil was supplied by ACH Food Company (Memphis, TN). Octadecane (approx 99%, EEC No 209-790-3, FW254.5) was purchased from Sigma Chemical Company (St Louis, MO). b-Lactoglobulin was a gift from Le Sueur Food Ingredient Company (Le Sueur, MN). Spray-dried sodium caseinate (ALANATE 180) with 1.2% sodium content and <0.1% calcium content was kindly provided by American Casein Company (Burlington, NJ). Tween 20 (P1379, CAS 9005-64-5), sodium dodecyl sulfate (SDS) (L-4509, EC No 205-788-1, FW 288.4), and sodium phosphate monobasic (S 0751, CAS 7558-80-7, FW 119.98) were purchased from the Sigma Chemical Company. Sodium phosphate dibasic anhydrous (S347-1, CAS 7558-79-4) was purchased from Fisher Scientific (Fair Lawn, NJ). Double distilled water was used for the preparation of all solutions. 2.2. Emulsion preparation Oil-in-water emulsions were prepared by homogenizing 5 wt% lipid phase (corn oil and/or octadecane) with 95 wt% aqueous phase (1e10 wt% emulsifier, 0e50 wt% glycerol, 10 mM sodium phosphate buffer, pH 7.0). A coarse emulsion premix was prepared by blending the lipid and aqueous phases together using a highspeed mixer (Bamix, Biospec Products, Bartlesville, OK) for 2 min at room temperature. Fine emulsions were formed by passing the coarse emulsions through an air-driven microfluidizer (Microfluidics, Newton, MA, USA). The coarse emulsions were fed into the microfluidizer through a 100 ml glass reservoir, and were passed through the homogenization unit for different numbers of passes (1e14) at various homogenization pressures (4e14 kbar). 2.3. Particle size measurements The particle size distribution and mean droplet diameter of diluted emulsions were measured by a commercial dynamic light scattering device (Nano-ZS, Malvern Instruments, Worcestershire, UK). Mean particle diameters were reported as “Z-average” diameters (the scattering intensity-weighted mean diameter), which were calculated from the signal intensity versus particle diameter data normalized to size increments using Mie Theory. Samples were diluted prior to making the particle size measurements to avoid multiple scattering effects with phosphate buffer (pH 7, no cosolvents) using a dilution factor of 1:100 sample-to-buffer. 2.4. Turbidity measurements The turbidity of selected samples was measured using a UVevisible spectrophotometer (Agilent 7010 Particle Size Analyzer, Agilent Technologies. Inc, CA, USA). The samples were diluted in buffer solution to a range of different oil droplet concentrations, and the turbidity was determined at 600 nm. 2.5. Viscosity measurements Viscosity measurements of selected samples were performed using a dynamic shear rheometer (Kinexus Rheometer, Malvern Instruments. Inc, MA, USA) using a shear rate profile from 0.1 to 1000 s
1. All the measurements were performed at 40  C.

1002

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008

2.6. Density measurements The densities of selected samples were measured using a digital oscillating tube method (DE50 Density Meter, Mettler-Toledo. Inc, OH, USA). The samples were degassed and filled into the density meter cell and allowed to equilibrate to 40  C. The density of the samples at 40  C was recorded after the sample temperature reached thermal equilibrium for at least 5 min. 2.7. Statistical analysis All measurements were performed on two or three freshly prepared samples and are reported as means and standard deviations. 3. Results and discussion 3.1. Impact of microfluidization pressure and number of cycles on particle size Initially, the effect of homogenization pressure and number of passes through the homogenizer on the mean particle size was determined for 5 wt% corn oil-in-water emulsions stabilized by 2 wt% b-lactoglobulin (Fig. 1). Overall, we observed a decrease in droplet diameter with increasing pressure and number of passes, which is in agreement with previous studies (Tan & Nakajima, 2005; Tcholakova, Denkov, Sidzhakova, Ivanov, & Campbell, 2003). There was a large decrease in the mean particle size when the coarse emulsions were passed once through the microfluidizer at all pressures studied. For example, the mean droplet diameter decreased from around 1130 nm before homogenization to 231, 214, 180, 177, 168 and 165 nm after a single pass at homogenization pressures of 4, 6, 8, 10, 12 and 14 kbar, respectively. These results indicate that the microfluidizer is particularly efficient at forming small droplet sizes in emulsions. The mean droplet diameter

continued to decrease as emulsions were passed through the homogenizer an increasing number of times, but the further reductions were fairly modest (Fig. 1). For example, at 14 kbar, the mean droplet diameter decreased from around 165 nm after a single pass to around 146 nm after eight passes. The mean droplet diameter did not change appreciably after about six passes, and therefore we used this value in the remainder of the experiments. The influence of homogenization pressure (P) on the mean droplet diameter (d) produced after six passes is shown in Fig. 2 for emulsions stabilized by 2 wt% b-lactoglobulin. The data is plotted as log diameter versus log pressure since previous studies suggest there should be a linear relationship providing there is sufficient emulsifier present (Walstra, 1993). The mean droplet diameter decreased from around 213 to 150 nm as the homogenization pressure was increased from 4 to 14 kbar. The decrease in droplet size with increasing pressure can be attributed to the increase in the magnitude of the disruptive forces generated within the homogenization chamber. However, a continued decrease in droplet size with increasing pressure need not be observed under all circumstances, e.g., if the amount of emulsifier present becomes limiting or if the emulsifier changes characteristics within the homogenizer (Jafari et al., 2008; Tcholakova et al., 2003). The minimum size of stable droplets that can be theoretically produced during homogenization given a certain amount of emulsifier is (McClements, 2005):

dmin ¼

6$G$f 6$G$f ¼ 0 cS cS ð1
fÞ

(1)

Here, G is the surface load of the emulsifier (in kg m
2), f is the disperse phase volume fraction, cS is the concentration of emulsifier 0 in the emulsion (in kg m
3) and cS is the concentration of emulsifier in the continuous phase (in kg m
3). In our study, the emulsions contained 5 wt% of oil droplets (f w 0.05) and 2 wt% of emulsifier 0 in the aqueous phase (cS w 20 kg m
3). The surface load of globular proteins at oil droplet surfaces has been reported to be around

1130 nm

2.34 4K

6K

8K

10K

12K

14K

2.32

Log (Mean Droplet Diameter / nm)

Mean Droplet Diameter (nm)

240

220

200

180

Log(d/nm) = -0.289 Log (P/kbar) + 2.499 R² = 0.9938

2.3 2.28 2.26 2.24 2.22 2.2

160

2.18 2.16 0.6

140 0

1

2

3

4

5

6

7

8

0.7

0.8

0.9

1

1.1

9

Number of passes Fig. 1. Effect of microfluidization pressure (in kbar) and number of passes on the mean droplet diameter produced in 5 wt% corn oil-in-water emulsions containing 2 wt% b-lactoglobulin.

Log (Pressure / kbar) Fig. 2. Effect of microfluidization pressure (in kbar) on the mean droplet diameter produced in 5 wt% corn oil-in-water emulsions containing 2 wt% b-lactoglobulin after 6 passes.

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008

3.2. Impact of emulsifier type and concentration on particle size A wide variety of different emulsifiers are available for use in the food industry, including small-molecule surfactants, phospholipids, proteins and polysaccharides (McClements, 2005; Stauffer, 1999; Whitehurst, 2006). It is therefore useful to establish the influence of emulsifier type on nanoemulsion formation using a microfluidizer. In this study, we examined the influence of four kinds of emulsifier on the droplet size: a globular protein (b-lactoglobulin), a flexible protein (sodium caseinate), a non-ionic surfactant (Tween 20), and an anionic surfactant (sodium dodecyl sulfate, SDS). 5 wt%

45

40

1 Pass

2 Passes

3 Passes

4 Passes

5 Passes

6 Passes

7 Passes

8 Passes

Relative Droplet Volume (%)

35

30

25

20

15

10

5

0 10

100

1000

10000

Droplet Diameter (nm) Fig. 3. Changes in particle size distribution of 5 wt% corn oil-in-water emulsions containing 2 wt% b-lactoglobulin with number of passes (at 14 kbar).

corn oil-in-water emulsions with different aqueous phase emulsifier concentrations (1e10 wt%) were prepared using fixed homogenization conditions (14 kbar, 6 passes), and then their particle sizes were measured (Fig. 5). The type and concentration of emulsifier present prior to homogenization clearly had a major impact on the size of the droplets produced. At 10 wt% emulsifier, which corresponds to

45

40

4k

6k

8k

10k

12k

14k

35

Relative Droplet Volume (%)

2 mg m
2 (G ¼ 2  10
6 kg m
2) (Tcholakova et al., 2003). Hence, the minimum droplet diameter that could theoretically have been achieved assuming that all of the protein adsorbed to the droplet surfaces during homogenization is about 32 nm. The mean droplet diameters produced in our study were appreciably higher than this value, which suggested that there was more than enough emulsifier initially present to cover all of the droplets formed by the microfluidizer. The fact that the observed minimum droplet diameter (w150 nm) achieved was considerably larger than this theoretical minimum value may have been because the emulsifier did not adsorb quickly enough during homogenization, or because the homogenizer was incapable of generating sufficiently intense disruptive forces at the pressures used. We found a linear decrease in log(d) with increasing log(P), with a slope of about
0.29 (Fig. 2). Mathematical analyses and computer simulations of the breakup of droplets within highpressure homogenizers suggest that the slope of log(d) versus log(P) should be around
0.6 when the fluid flow conditions are primarily turbulent-inertial (Hakansson, Tragardh, & Bergenstahl, 2009; Walstra, 2003). This suggests that either droplet breakup was not primarily due to turbulent-inertial forces in the bench-scale microfluidizer used in this study, or that there was retardation of droplet breakup or significant re-coalescence at higher pressures (Jafari et al., 2008). For comparison, we therefore measured the decrease in log(d) with increasing log(P) for a similar emulsion containing 2 wt% SDS. SDS is a small-molecule surfactant that is believed to rapidly adsorb to oil droplet surfaces within homogenizers (Schubert & Engel, 2004). For this system, the mean droplet diameter decreased from around 245 to 123 nm as the homogenization pressure was increased from 4 to 14 kbar, with the slope of log(d) versus log(P) being around
0.57 (Fig. 2), which is close to the value reported for turbulent-inertial breakup (Hakansson et al., 2009; Walstra, 2003). We therefore attribute the relatively low value of the slope for b-lactoglobulin to its relatively slow adsorption rate leading to some re-coalescence, or due to its ability to form a protective film around the lipid droplets that limited their further disruption. Further insight into this phenomenon could be obtained in future studies by measuring the droplet re-coalescence rate within the homogenizer using previously established methods (Jafari et al., 2008). The overall particle size distribution, rather than just the mean particle diameter, is important for many practical applications of nanoemulsions. We therefore plotted the change in particle size distribution with number of passes and pressures for selected blactoglobulin-stabilized samples (Figs. 3 and 4). This data shows that the width of the particle size distribution decreased appreciably with increasing number of passes (Fig. 3) and increasing pressure (Fig. 4). There may therefore be an advantage for certain applications to use higher number of passes or higher homogenization pressures to obtain narrower particle size distributions and decrease the fraction of large droplets present. For example, to prevent the formation of a visible ring at the top of a beverage emulsion due to rapid creaming of a population of large droplets.

1003

30

25

20

15

10

5

0 10

100

1000

10000

Droplet Diameter (nm) Fig. 4. Changes in particle size distribution of 5 wt% corn oil-in-water emulsions containing 2 wt% b-lactoglobulin with homogenization pressure (in kbar) for six passes.

1004

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008

220 Tween20 SDS

200

Casein

Mean Droplet Diameter (nm)

BLG

180

160

140

phase that contributed to the light scattering signal. In contrast to b-lactoglobulin, caseinate has a more disordered and flexible structure that is not susceptible to dramatic changes in its surface activity or aggregation behavior due to denaturation. This would explain why there was a decrease in droplet diameter with increasing pressure for caseinate, but an increase for b-lactoglobulin (Fig. 5). It should be noted that in practice it is often desirable to minimize the total amount of emulsifier used to prepare an emulsion due to economic, sensory, or regulatory reasons. For example, many small-molecule surfactants generate off-flavors when used at sufficiently high concentrations. For beverage applications this is less of a problem because the initial concentrated emulsions prepared by homogenization are usually diluted considerably when they are incorporated into final beverage products. 3.3. Impact of phase viscosities on particle size

120

100 0

2

4

6

8

10

Emulsifier Concentration (wt%) Fig. 5. Influence of emulsifier type and concentration on mean particle diameter for 5 wt% corn oil-in-water emulsions (pH 7, 40  C) prepared at 14 kbar using 6 passes.

a 2:1 ratio of emulsifier-to-oil, the mean droplet diameters produced were 110, 113, 179 and 162 nm for SDS, Tween 20, sodium caseinate, and b-lactoglobulin, respectively. This result supports previous work that has also shown that small-molecule surfactants (like SDS and Tween) are more effective at making small droplets than biopolymers (caseinate and b-lactoglobulin) under similar homogenization conditions in high-pressure homogenizers because they adsorb to the droplet surfaces more rapidly (Karbstein & Schubert, 1995; Stang, Karbstein, & Schubert, 1994). For the small-molecule surfactants and caseinate there was a decrease in mean droplet diameter with increasing emulsifier concentration (Fig. 5). This trend might be expected because there should be more emulsifier present to cover any new droplet surfaces formed during homogenization, and because the droplet surfaces will be covered more rapidly by a layer of emulsifier molecules. Nevertheless, the decrease in diameter was appreciably less than might be expected from Equation (1), which predicts that the minimum mean droplet diameter should be inversely proportional to the surfactant concentration. For example, when the SDS concentration was increased from 1 to 10 wt%, the droplet diameter only decreased from 131 to 110 nm, whereas one might expect a 10-fold reduction. This suggests that the droplet size was mainly limited by the maximum disruptive forces that could be generated within the homogenizer at the operating pressure used, rather than by the amount of emulsifier present. Interestingly, there was actually an increase in the mean particle diameter with increasing emulsifier content for the emulsions containing b-lactoglobulin (Fig. 5). One possible explanation for this phenomenon is that some of this globular protein became denatured during the high-pressure homogenization process (Rampon, Riaublanc, Anton, Genot, & McClements, 2003). At sufficiently high concentrations b-lactoglobulin denaturation could lead to an increase in measured droplet size through a number of mechanisms: (i) unfolded proteins may have formed multilayers around each droplet; (ii) unfolded proteins may have promoted droplet flocculation by increasing droplet surface hydrophobicity; (iii) unfolded proteins may have formed protein aggregates in the continuous

The rheological properties of the oil and aqueous phases used to prepare oil-in-water emulsions in the food and beverage industries may vary considerably depending on their composition. For example, flavor oils tend to have much lower viscosities than triglyceride oils, while addition of solutes (such as sugars, polyols, and biopolymers) to aqueous solutions may greatly increase their viscosities. Previous theoretical and experimental studies have shown that phase viscosities alter the efficiency of droplet disruption, and therefore the minimum droplet size attainable, within various types of homogenizers (Jafari et al., 2008; Walstra, 1993, 2003). It is therefore important to establish the influence of the dispersed and continuous phase viscosities on the formation of nanoemulsions. This knowledge can then be used to optimize the conditions required to form nanoemulsions using the microfluidization approach. 3.3.1. Impact of dispersed phase viscosity A series of 5 wt% oil-in-water emulsions was prepared with similar aqueous phase compositions, but different oil phase compositions using fixed homogenization conditions (14 kbar, 6 passes). The viscosity of the disperse phase was varied by using different combinations of corn oil and octadecane as the oil phase. Corn oil has a higher viscosity than octadecane, and so an increase in octadecane concentration corresponds to a decrease in oil phase viscosity (Table 1a). The viscosity of the continuous phase was kept constant by using a fixed aqueous phase composition (phosphate buffer). The influence of two different types of emulsifiers was examined: a small-molecule ionic surfactant (2.5 wt% SDS) and a globular protein (2.5 wt% b-lactoglobulin). The temperature was maintained at about 40  C during homogenization for all the emulsions to ensure that it was well above the crystallization temperature of octadecane (28  C). When SDS was used as the emulsifier, there was an appreciable increase in the mean droplet diameter with increasing corn oil concentration in the oil phase e.g., d increased from around 92 nm for pure octadecane to around 125 nm for pure corn oil (Fig. 6). On Table 1a Physicochemical properties of oil phases with different compositions at 40  C: corn oil ¼ CO; octadecane ¼ O. Oil composition

Density (kg m
3)

Viscosity (mPa s)

0% CO:100% 20% CO:80% 40% CO:60% 60% CO:40% 80% CO:20% 100% CO:0%

774.1  0.1 798.5  0.1 824.2  0.1 851.6  0.1 869.2  0.1 912.7  0.1

5  1.2 7.6  1.3 11  0.7 13.2  0.8 21  2.1 31.4  2.2

O O O O O O

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008

around the oil droplets that was more resistant to disruption within the homogenizer, therefore retarding droplet breakup within the disruption zone. Second, the b-lactoglobulin may have adsorbed more slowly to the droplet surfaces than SDS and therefore more droplet re-coalescence may have occurred within the homogenizer. Our results are in agreement with a recent study using a microfluidizer that showed that much smaller droplets could be produced in oil-in-water emulsions when low viscosity alkane oils were used as the oil phase rather than high viscosity triglyceride oils (Wooster et al., 2008).

160

150

140

Mean Droplet Diameter (nm)

1005

130

120

110

100

BLG SDS

90

80 0

20

40

60

80

100

Octadecane in Oil Phase (wt%) Fig. 6. Influence of oil phase concentration (octadecane:corn oil) on mean particle diameter for 5 wt% oil-in-water emulsions (pH 7, 40  C) prepared at 14 kbar using 6 passes for systems stabilized by 2 wt% b-lactoglobulin or SDS.

the other hand, when b-lactoglobulin was used as the emulsifier, there was little change in the mean droplet diameter with increasing corn oil concentration (Fig. 6). These results indicate that both surfactant type and oil composition had a major influence on the minimum droplet size that could be produced. Previous studies have also shown that droplet breakup becomes more difficult as the viscosity of the disperse phase increases for various types of homogenization device (Jafari et al., 2008; Seekkuarachchi, Tanaka, & Kumazawa, 2006; Walstra, 1993). The time for the droplets to become deformed within the disruption zone of a homogenizer increases as the disperse phase viscosity increases (Hakansson et al., 2009; Walstra, 1993). Droplets tend to be disrupted when the residence time within the disruption zone is longer than their deformation time (Walstra, 1993). Hence, more viscous oil droplets are more difficult to breakup within a homogenizer because they may leave the disruption zone before they have had time to become deformed and disrupted. It has been proposed that the dependence of mean droplet diameter (d) on disperse phase viscosity (hD) can be described by the following empirical expression (Jafari et al., 2008):

d ¼ AhbD

(2)

where A and b are empirically determined constants, with b been reported to lie between 0.2 and 0.9 for high-pressure homogenizers depending on the major disruption mechanism involved (Schultz, Wagner, Urban, & Ulrich, 2004). The values in Table 1a were used to plot log(d/nm) versus log(hD/mPa s) for the emulsions stabilized by either SDS or b-lactoglobulin. We found that b was around 0.17 when SDS was used as the emulsifier, and around 0.016 when blactoglobulin was used. These values are less than those reported by Schultz and co-workers, which may have been at least partly because the interfacial tension increases with increasing octadecane concentration, which would oppose droplet disruption (see later). The observation that the droplet diameter was much less dependent on disperse phase viscosity for b-lactoglobulin than for SDS emulsions may be due to a number of physicochemical phenomena. First, the b-lactoglobulin may have formed a coating

3.3.2. Impact of continuous phase viscosity A series of 5 wt% oil-in-water emulsions was prepared with similar oil phase compositions, but different aqueous phase compositions using fixed homogenization conditions (14 kbar, 6 passes). The viscosity of the disperse phase was kept constant by using pure corn oil as the oil phase. The viscosity of the continuous phase was varied by including different amounts of glycerol (0e50 wt%) in the aqueous phase prior to homogenization. Glycerol has a much higher viscosity than water, and so the aqueous phase viscosity increases with increasing glycerol content (Table 1b). Again we compared the effect of using two different types of emulsifiers (2.5 wt% SDS or b-lactoglobulin) on the droplet size produced by the microfluidizer. When SDS was used as the emulsifier, there was an appreciable reduction in the mean droplet diameter with increasing glycerol concentration e.g., d decreased from around 127 nm for 0% glycerol to around 87 nm for 50% glycerol (Fig. 7). On the other hand, when b-lactoglobulin was used as the emulsifier, there was only a slight decrease in the mean droplet diameter with increasing glycerol concentration e.g., d decreased from around 150 nm for 0% glycerol to around 142 nm for 50% glycerol (Fig. 7). The greater sensitivity of droplet size to continuous phase viscosity for SDS than for b-lactoglobulin was therefore similar to that found for the dependence on disperse phase viscosity (Section 3.3.1). It has been proposed that an expression similar to Equation (2) can also be used to describe the dependence of mean droplet diameter (d) on disperse phase viscosity (hC):

d ¼ A0 hb’ C

(3)

A0

b0

where and are empirically determined constants. In this case, we found that b0 was around
0.24 when SDS was used as the emulsifier, and around
0.029 when b-lactoglobulin was used. Again, the SDS-stabilized emulsions are much more sensitive to phase viscosity than the b-lactoglobulin-stabilized ones, which can be attributed to the physicochemical mechanisms discussed above. Potentially, changes in continuous phase viscosity may influence droplet size through a number of different mechanisms: (i) increased droplet fragmentation due to increased disruptive shear stresses; (ii) decreased droplet re-coalescence due to decreased

Table 1b Physicochemical properties of aqueous phases with different compositions at 40  C: glycerol ¼ G; water ¼ W. The densities were measured in our laboratory while the viscosity values are published values (www.dow.com/glycerine/resources/ physicalprop.htm). Aqueous phase composition

Density (kg m
3)

Viscosity (mPa s)

0% G:100% 10% G:90% 30% G:80% 30% G:70% 40% G:60% 50% G:50%

992.5  0.1 1015.9  0.1 1039.7  0.1 1064.4  0.1 1090.0  0.2 1116.2  0.2

0.656 0.826 1.07 1.46 2.07 3.1

W W W W W W

1006

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008 160

150

Mean Droplet Diameter (nm)

140

130

120

110

100 BLG

90

SDS

80 0

10

20

30

40

50

Glycerol in Aqueous Phase (wt%) Fig. 7. Influence of aqueous phase concentration (glycerol:water) on mean particle diameter for 5 wt% oil-in-water emulsions (pH 7, 40  C) prepared at 14 kbar using 6 passes for systems stabilized by 2 wt% b-lactoglobulin or SDS.

droplet collision frequency; (iii) increased droplet re-coalescence due to a reduction in emulsifier adsorption rate (McClements, 2005; Walstra, 1993). The relative importance of these factors is highly system dependent, and depends on homogenizer design and operating conditions (Hakansson et al., 2009). In homogenizers where droplet disruption is primarily due to turbulence, droplet size is not believed to be strongly dependent on continuous phase viscosity, i.e., b0 <
0.03 (Schultz et al., 2004). On the other hand, in homogenizers where shear forces play an important role in droplet disruption, the droplet diameter should decrease with increasing continuous phase viscosity (Hakansson et al., 2009). The appreciable decrease in droplet diameter observed with increasing continuous phase viscosity for the SDS emulsions in this work therefore suggests that shear forces did play an important role in droplet disruption and/or coalescence in the laboratory scale microfluidizer used. Our study is in agreement with recent work by Wooster and co-workers who found that addition of polyethylene glycol (PEG) to the aqueous phase of oil-in-water emulsions prior to microfluidization led to a decrease in mean droplet diameter, which was partially attributed to the ability of PEG to increase the continuous phase viscosity (Wooster et al., 2008). 3.3.3. Impact of viscosity ratio on particle size Practically, it is often possible to control both the continuous phase and disperse phase viscosities so as to minimize the droplet size produced during homogenization. Previous studies suggest that there is an optimum range of disperse-to-continuous phase viscosity ratios (0.1 < hD/hC < 5) for producing small droplets in various kinds of homogenizers (Schubert, Ax, & Behrend, 2003; Schubert & Engel, 2004; Walstra, 1993, 2003; Wooster et al., 2008). We therefore plotted the mean droplet diameters versus viscosity ratio for the various emulsions studied (Fig. 8). In addition to the data for the emulsions shown in Figs. 6 and 7, we also included data for a series of emulsions that had different oil phase compositions (0%, 50% or 100% corn oil) as well as aqueous phase compositions (0e50 wt% glycerol). All the systems studied were 5 wt% oil-in-water emulsions stabilized by 2.5 wt% emulsifier prepared using fixed homogenization conditions (14 kbar, 6 passes, 40  C).

Fig. 8. Influence of ratio of oil phase to aqueous phase viscosity on the mean particle diameter for 5 wt% oil-in-water emulsions (pH 7, 40  C) prepared at 14 kbar using 6 passes for systems stabilized by 2 wt% b-lactoglobulin or SDS.

The mean droplet diameter plot increased with increasing viscosity ratio for both types of emulsifiers used, but the droplet sizes were always higher for b-lactoglobulin-emulsions than for SDS emulsions at the same viscosity ratio (Fig. 8). In addition, the change in mean diameter with viscosity ratio was also different for the two emulsions. For SDS emulsions, the slope of log(d) versus log (hD/hC) was around 0.23, while for b-lactoglobulin-emulsions it was around 0.00. These results again highlight that smaller droplets can be produced using a small-molecule surfactant than a protein under similar homogenization conditions. The smallest droplets (d ¼ 57 nm) that could be produced were for the SDS emulsions containing 50 wt% glycerol in the aqueous phase and 100% octadecane in the oil phase since this combination minimized the viscosity ratio. When b-lactoglobulin-emulsions were prepared using similar aqueous and oil phase compositions the mean droplet diameter was appreciably larger (d ¼ 151 1 nm). At this droplet size the emulsions still had an appreciable turbidity. Hence, this study shows that it is difficult to prepare globular protein-stabilized nanoemulsions containing droplets small enough to only scatter light weakly. 3.3.4. Impact of interfacial tension Finally, it should be noted that other factors may also influence the size of the droplets produced during homogenization. In particular, the efficiency of droplet disruption depends on the oilewater interfacial tension (g), with the droplet size produced during homogenization decreasing with decreasing interfacial tension (Schubert & Engel, 2004; Walstra, 1993). The interfacial tension at an oil phaseeaqueous phase system depends on the composition of the two immiscible liquids. The interfacial tension of aqueous glycerol solutions only decreases slightly with increasing glycerol concentration, e.g., it has been reported that g decreased less than 10% when the glycerol level was increased from 0 to 50 wt% (Chanasattru, Decker, & McClements, 2007). One would therefore expect the change in interfacial tension due to varying

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008

glycerol concentrations to have little impact on droplet disruption and size. On the other hand, the interfacial tensions of corn oile water and octadecaneewater systems have been reported to be around 31.5 mN m
1 (Chaiyasit, McClements, Weiss, & Decker, 2008) and 52 mN m
1 (Hiemenz & Rajagopalan, 1997), respectively. The fact that the interfacial tension is considerably higher for octadecane than for corn oil suggests that the Laplace pressure opposing droplet deformation and disruption would be higher for octadecane than for corn oil droplets (McClements, 2005). This phenomenon may account for the fact that the slope (b) of log(d) versus log(hD) for the emulsions used in this work (Fig. 6) was somewhat less than those reported by previous workers (Schultz et al., 2004). The decrease in viscosity with increasing octadecane concentration favored droplet disruption, whereas the increase in interfacial tension opposed it. 3.4. Optical properties of emulsions For many practical applications the optical properties of emulsion-based delivery systems is particularly important. In some applications it is important to have an optically transparent delivery system (e.g., clear beverages), whereas in others this may be unimportant (e.g., in an opaque food) or it may even be desirable to have a turbid delivery system (e.g., to impart cloudiness to some soft drinks). The optical properties of selected emulsions were therefore studied. We used 5 wt% octadecane oil-in-water emulsions stabilized by 2.5 wt% SDS since these were previously shown to give the smallest droplet diameters. Emulsions were prepared containing either 0 or 50 wt% glycerol in the aqueous phase during homogenization to obtain systems with different droplet sizes. Prior to dilution, the emulsion containing 0% glycerol was optically opaque whereas the emulsion containing 50% glycerol was slightly turbid, which may have been because of differences in droplet size or due to the fact that glycerol reduces the refractive index contrast between the oil and aqueous phases (McClements, 2002b). We therefore measured the variation in the turbidity (at 600 nm) of the emulsions after dilution as a function of droplet concentration (Fig. 9). These emulsions were diluted with buffer

0.2 0% Glycerol

0.18

τ = 0.462c + 0.0023 R² = 0.9969

50%glycerol

solution (at least 1:10) so that the final glycerol concentration (<5%) would not be expected to significantly affect the refractive index of the aqueous phase and therefore the overall light scattering characteristics of the emulsions. The increase in turbidity with increasing droplet concentration was much higher for the emulsion that contained 0% glycerol during homogenization (0.46 cm
1%
1) than the one containing 50% glycerol during homogenization (0.15 cm
1%
1), which can be attributed to differences in droplet size: d ¼ 95 and 57 nm for 0 and 50% glycerol, respectively. Smaller droplets are known to scatter light less efficiently than larger ones in this particle size range, which accounts for the lower turbidities of the emulsions containing smaller droplet sizes (McClements, 2002b). This study shows that emulsions that do not scatter light strongly can be produced by homogenizing a relatively low viscosity oil phase with a relatively high viscosity aqueous phase. In the food industry, flavor oils or essential oils could be used as low viscosity oils, rather than hydrocarbons. 4. Conclusions This study examined the influence of emulsion composition and homogenization conditions on the size of the oil droplets produced using a microfluidizer. Nanoemulsions (d < 100 nm) that only scattered light weakly could be produced by optimizing homogenization conditions (pressure and number of passes), disperse and aqueous phase viscosities, and emulsifier type. Small-molecule surfactants (Tween 20 and SDS) were found to produce smaller droplet sizes than biopolymers (b-lactoglobulin or caseinate), which was attributed to differences in adsorption rates and interfacial properties. In this study, an alkane oil (octadecane) was used to decrease the viscosity of the oil phase and produce smaller droplets, but other types of food-grade oils with low viscosities could also be used, such as flavor oils or essential oils. We also showed that addition of relatively high concentrations of watersoluble cosolvents (such as glycerol) to the aqueous phase prior to homogenization can also be used to reduce the droplet sizes. This study has important implications for optimizing the composition and homogenization conditions required to produce food-grade nanoemulsions. Acknowledgements

0.16

This material is partly based upon work supported by United States Department of Agriculture, CREES, NRI Grants, and Massachusetts Department of Agricultural Resources CTAGR7AGI UMA 00 Grant. We also acknowledge funding from the University of Massachusetts (CVIP and Hatch).

0.14

Turbidity (cm-1)

1007

0.12 0.1

References 0.08 0.06 0.04 τ = 0.153c + 0.0005 R² = 0.9922

0.02 0 0

0.1

0.2

0.3

0.4

Droplet Concentration (wt%) Fig. 9. Impact of glycerol content on turbidity of 5 wt% octadecane oil-in-water emulsions diluted to different droplet concentrations (2.5 wt% SDS, 40  C, 14 kbar, 6 passes).

Acosta, E. (2009). Bioavailability of nanoparticles in nutrient and nutraceutical delivery. Current Opinion in Colloid & Interface Science, 14(1), 3e15. Anton, N., Benoit, J. P., & Saulnier, P. (2008). Design and production of nanoparticles formulated from nano-emulsion templates e a review. Journal of Controlled Release, 128(3), 185e199. Azeem, A., Rizwan, M., Ahmad, F. J., Khar, R. K., Iqbal, Z., & Talegaonkar, S. (2009a). Components screening and influence of surfactant and cosurfactant on nanoemulsion formation. Current Nanoscience, 5(2), 220e226. Azeem, A., Rizwan, M., Ahmad, F. J., Iqbal, Z., Khar, R. K., Aqil, M., et al. (2009b). Nanoemulsion components screening and selection: a technical note. AAPS PharmSciTech, 10(1), 69e76. Bouchemal, K., Briancon, S., Perrier, E., & Fessi, H. (2004). Nano-emulsion formulation using spontaneous emulsification: solvent, oil and surfactant optimisation. International Journal of Pharmaceutics, 280(1e2), 241e251. Chaiyasit, W., McClements, D. J., Weiss, J., & Decker, E. A. (2008). Impact of surfaceactive compounds on physicochemical and oxidative properties of edible oil. Journal of Agricultural and Food Chemistry, 56(2), 550e556.

1008

C. Qian, D.J. McClements / Food Hydrocolloids 25 (2011) 1000e1008

Chanasattru, W., Decker, E. A., & McClements, D. J. (2007). Physicochemical basis for cosolvent modulation of beta-lactoglobulin functionality: interfacial tension study. Food Research International, 40(8), 1098e1105. Chu, B. S., Ichikawa, S., Kanafusa, S., & Nakajima, M. (2007). Preparation and characterization of beta-carotene nanodispersions prepared by solvent displacement technique. Journal of Agricultural and Food Chemistry, 55(16), 6754e6760. Freitas, S., Merkle, H. P., & Gander, B. (2005). Microencapsulation by solvent extraction/evaporation: reviewing the state of the art of microsphere preparation process technology. Journal of Controlled Release, 102(2), 313e332. Gutierrez, J. M., Gonzalez, C., Maestro, A., Sole, I., Pey, C. M., & Nolla, J. (2008). Nanoemulsions: new applications and optimization of their preparation. Current Opinion in Colloid & Interface Science, 13(4), 245e251. Hakansson, A., Tragardh, C., & Bergenstahl, B. (2009). Dynamic simulation of emulsion formation in a high pressure homogenizer. Chemical Engineering Science, 64(12), 2915e2925. Hiemenz, P. C., & Rajagopalan, R. (1997). Principles of colloid and surface chemistry (3rd ed.). New York: Marcel Dekker. Hu, J. H., Johnston, K. P., & Williams, R. O. (2004). Nanoparticle engineering processes for enhancing the dissolution rates of poorly water soluble drugs. Drug Development and Industrial Pharmacy, 30(3), 233e245. Huang, Q. R., Yu, H. L., & Ru, Q. M. (2010). Bioavailability and delivery of nutraceuticals using nanotechnology. Journal of Food Science, 75(1), R50eR57. Jafari, S. M., He, Y. H., & Bhandari, B. (2006). Nano-emulsion production by sonication and microfluidization e a comparison. International Journal of Food Properties, 9(3), 475e485. Jafari, S. M., He, Y. H., & Bhandari, B. (2007). Production of sub-micron emulsions by ultrasound and microfluidization techniques. Journal of Food Engineering, 82(4), 478e488. Jafari, S. M., Assadpoor, E., He, Y. H., & Bhandari, B. (2008). Re-coalescence of emulsion droplets during high-energy emulsification. Food Hydrocolloids, 22(7), 1191e1202. Karbstein, H., & Schubert, H. (1995). Developments in the continuous mechanical production of oil-in-water macro-emulsions. Chemical Engineering and Processing, 34(3), 205e211. Kesisoglou, F., Panmai, S., & Wu, Y. H. (2007). Application of nanoparticles in oral delivery of immediate release formulations. Current Nanoscience, 3(2), 183e190. Leong, T. S. H., et al. (2009). Minimising oil droplet size using ultrasonic emulsification. Ultrasonics Sonochemistry, 16(6), 721e727. Lesmes, U., & McClements, D. J. (2009). Structureefunction relationships to guide rational design and fabrication of particulate food delivery systems. Trends in Food Science & Technology, 20(10), 448e457. McClements, D. J. (2002a). Colloidal basis of emulsion color. Current Opinion in Colloid & Interface Science, 7(5e6), 451e455. McClements, D. J. (2002b). Theoretical prediction of emulsion color. Advances in Colloid and Interface Science, 97(1e3), 63e89. McClements, D. J. (2005). Food emulsions: Principles, practice, and techniques (2nd ed.). In CRC series in contemporary food science Boca Raton: CRC Press. McClements, D. J. (2010). Emulsion design to improve the delivery of functional lipophilic components. Annual Review of Food Science and Technology, 1(1), 241e269. McClements, D. J., Decker, E. A., & Weiss, J. (2007). Emulsion-based delivery systems for lipophilic bioactive components. Journal of Food Science, 72(8), R109eR124.

Perrier-Cornet, J. M., Marie, P., & Gervais, P. (2005). Comparison of emulsification efficiency of protein-stabilized oil-in-water emulsions using jet, high pressure and colloid mill homogenization. Journal of Food Engineering, 66(2), 211e217. Rampon, V., Riaublanc, A., Anton, M., Genot, C., & McClements, D. J. (2003). Evidence that homogenization of BSA-stabilized hexadecane-in-water emulsions induces structure modification of the nonadsorbed protein. Journal of Agricultural and Food Chemistry, 51(20), 5900e5905. Sanguansri, P., & Augustin, M. A. (2006). Nanoscale materials development e a food industry perspective. Trends in Food Science & Technology, 17(10), 547e556. Schubert, H., Ax, K., & Behrend, O. (2003). Product engineering of dispersed systems. Trends in Food Science & Technology, 14(1), 9e16. Schubert, H., & Engel, R. (2004). Product and formulation engineering of emulsions. Chemical Engineering Research & Design, 82(A9), 1137e1143. Schultz, S., Wagner, G., Urban, K., & Ulrich, J. (2004). High-pressure homogenization as a process for emulsion formation. Chemical Engineering & Technology, 27(4), 361e368. Seekkuarachchi, I. N., Tanaka, K., & Kumazawa, H. (2006). Formation and characterization of submicrometer oil-in-water (O/W) emulsions, using high-energy emulsification. Industrial & Engineering Chemistry Research, 45(1), 372e390. Stang, M., Karbstein, H., & Schubert, H. (1994). Adsorption-kinetics of emulsifiers at oilewater interfaces and their effect on mechanical emulsification. Chemical Engineering and Processing, 33(5), 307e311. Stauffer, S. (1999). Emulsifiers. St Paul, MN: Eagan Press. Tadros, T., Izquierdo, R., Esquena, J., & Solans, C. (2004). Formation and stability of nano-emulsions. Advances in Colloid and Interface Science, 108e109, 303e318. Tan, C. P., & Nakajima, M. (2005). beta-Carotene nanodispersions: preparation, characterization and stability evaluation. Food Chemistry, 92(4), 661e671. Tcholakova, S., Denkov, N. D., Sidzhakova, D., Ivanov, I. B., & Campbell, B. (2003). Interrelation between drop size and protein adsorption at various emulsification conditions. Langmuir, 19(14), 5640e5649. Velikov, K. P., & Pelan, E. (2008). Colloidal delivery systems for micronutrients and nutraceuticals. Soft Matter, 4(10), 1964e1980. Walstra, P. (1993). Principles of emulsion formation. Chemical Engineering Science, 48, 333. Walstra, P. (2003). Physical chemistry of foods. New York, NY: Marcel Dekker. Weiss, Decker, McClements, Kristbergsson, Helgason, & Awad. (2008). Solid lipid nanoparticles as delivery systems for bioactive food components. Food Biophysics, 3(2), 146e154. Whitehurst, R. J. (2006). Emulsifiers in food technology. Oxford, UK: Blackwell Publishing. Wissing, S. A., Kayser, O., & Muller, R. H. (2004). Solid lipid nanoparticles for parenteral drug delivery. Advanced Drug Delivery Reviews, 56(9), 1257e1272. Wooster, T. J., Golding, M., & Sanguansri, P. (2008). impact of oil type on nanoemulsion formation and Ostwald ripening stability. Langmuir, 24(22), 12758e12765. Yin, Chu, Kobayashi, & Nakajima (2008). Performance of selected emulsifiers and their combinations in the preparation of beta-carotene nanodispersions. In 9th International hydrocolloids conference, Singapore. Yuan, Y., Gao, Y. X., Zhao, J., & Mao, L. (2008). Characterization and stability evaluation of beta-carotene nanoemulsions prepared by high pressure homogenization under various emulsifying conditions. Food Research International, 41(1), 61e68.