Journal of Food Engineering xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: [Link]/locate/jfoodeng

Influence of surfactant and processing conditions in the stability

of oil-in-water nanoemulsions
Hélder D. Silva ⇑, Miguel A. Cerqueira, António A. Vicente
Centre of Biological Engineering, Universidade do Minho, Campus de Gualtar, 4710-057 Braga, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: This work evaluates the influence of the type of surfactant (Tween 20, SDS and DTAB) and processing con-
Received 14 January 2015 ditions on the stability of oil-in-water nanoemulsions, measured in terms of hydrodynamic diameter
Received in revised form 9 June 2015 (Hd), polydispersity index (PdI) and zeta potential (Zp). Nanoemulsions were prepared using
Accepted 18 July 2015
high-pressure homogenization based on a 24 level factorial design. Results show that processing param-
Available online xxxx
eters such as homogenization pressure, surfactant concentrations and oil:water ratio significantly
affected the values of Hd and PdI of nanoemulsions. The value of Hd of anionic nanoemulsions decreased
Keywords:
(from 177 to 128 nm) with the increase of the homogenization pressure. The increase in the surfactant
Nanoemulsion
Process conditions
concentration and the decrease of the oil:water ratio lead to a decrease of Hd for the cationic nanoemul-
Physical characterization sions (from 198 to 135 nm). The increase of the oil:water ratio lead to a decrease of Hd for the non-ionic
Medium chain triglycerides nanoemulsions (from 341 to 171 nm); this is contrary to the usual assumption that higher content in oil
Kinetic stability results in higher values of Hd. Those nanoemulsions showed a good kinetic stability (evaluated after cen-
trifugation, heating–cooling cycles and thermal stress) upon measuring the Hd during 28 and 35 days of
storage, without visual evidence of creaming and phase separation. After one year of storage the
nanoemulsions produced with the anionic surfactant remained kinetically stable, without visual evidence
of creaming and/or phase separation.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction or composition (Komaiko and McClements, 2015; Silva et al.,

2012; Solans and Solé, 2012). High-energy methods make use of
Driven by consumers’ demands for new and healthier food devices that apply high mechanical energy inputs to disrupt and
products the food industry seeks for new methodologies able to combine the oil and water phases, forming small droplets (Abbas
encapsulate, protect and release functional compounds. Based on et al., 2013; Cerqueira et al., 2014; Komaiko and McClements,
this, researchers are focusing their efforts in relevant issues to food 2015; Silva et al., 2012). High-pressure homogenization is pointed
and nutrition regarding the improvement of food quality, through as the most appropriate method for industrial applications, due to
nanotechnology (Cerqueira et al., 2013; Silva et al., 2012). the facility of operation, scalability, reproducibility, and high
Nanoemulsions are interesting for the food industry due to their throughput (Cerqueira et al., 2014; Donsì et al., 2011).
potential applications as delivery systems of bioactive compounds Preparing a nanoemulsion by a high-energy method implies
while preventing their degradation and improving their bioavail- using an oily and an aqueous phase, a surfactant and energy
ability (Donsì et al., 2011; Guttoff et al., 2015; Silva et al., 2012). (Tadros et al., 2004; Walstra, 1993). The nano emulsification pro-
Nanoemulsions generally consist of lipid droplets between 10 cess by high-pressure homogenization comprises both the defor-
and 200 nm dispersed in an aqueous phase, where surfactant mation and disruption of the droplets with the subsequent
molecules surround each oil droplet (Acosta, 2009; Cerqueira increase of the surface area and at the same time the droplet stabi-
et al., 2014). Nanoemulsions can be produced using either lization occurs by means of adsorption of the emulsifiers at the
low-energy or high-energy methods. Low-energy methods mainly interface of the droplets (Donsì et al., 2011; Stang et al., 2001).
depend on the intrinsic physicochemical properties of surfactants The difference in the interfacial free energy between the initial
and oily phase, forming nanoemulsions by simple mixing proce- and final state is by definition equal to the increase on the surface
dures or by changing the system conditions such as temperature area between the oily and aqueous phases multiplied by the inter-
facial tension (McClements, 2005). The energy necessary to
⇑ Corresponding author. increase the interfacial area (cDA) is very high and positive (i.e. it
E-mail address: heldersilva@[Link] (H.D. Silva). increases after homogenization), while the small entropy of the

[Link]
0260-8774/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
2 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx

dispersion, and the corresponding energy (TDS) is also positive but and SunActive Iron beverages, from NutraLease, AquaNova and
cannot compensate the interfacial free energy (McClements, 2005; High Vive company’s, respectively (AquaNova, 2013a; High Vive,
Schramm, 2006b; Tadros et al., 2004) and therefore Eq. (1) is 2013; NutraLease, 2011); food colorants, with NovaSol BCS and
always positive. Color Emulsion from AquaNova and Wild Flavours Inc, respectively
(AquaNova, 2013b; Wild Flavors Inc., 2013); food packaging with
DGformation ¼ cDA  T DS ð1Þ
BioNutriCoat from Improveat (Improveat, 2014); food supple-
Thus nanoemulsion formation is always thermodynamically ments, with NanoResveratrol from Life Enhancement (Life
unfavorable, due to the increase of the interfacial area after emul- Enhancement, 2013), Spray for Life from NanoSinergy
sification and to the energy required to produce the droplets (NanoSinergy, 2013) and NutriNano CoQ-10 from Solgar (Solgar,
(McClements, 2005; Tadros et al., 2004; Walstra, 1993). In order 2013); and fortified oils with Canova Active Oil from Shemen
to break up a droplet into smaller ones, it must be strongly Industries (Shemen Industries, 2013).
deformed and this is opposed by the Laplace pressure, Dp, which One of the aims of this work was to study the effects of differ-
is the difference in pressure between the inside and outside of ent charge surfactants, an anionic surfactant, Sodium Dodecyl
the droplet, being the pressure greater on the inside of the droplet, Sulphate (SDS); a cationic surfactant, dodecyltrimethylammo-
given by: nium bromide (DTAB) and a non-ionic surfactant, Tween 20
(Pinheiro et al., 2013). The effect of the process conditions i.e.,
Dp ¼ cð1=R1 þ 1=R2 Þ ð2Þ
pressure, number of cycles, surfactant concentrations and oil con-
where R1 and R2 are the principal radii of curvature of the droplet. tent in the mean hydrodynamic diameter (Hd), polydispersity
For a nanoemulsion, a spherical droplet yields R1 = R2 = R index (PdI), zeta potential (Zp) and the stability of the nanoemul-
(Schramm, 2006b; Tadros et al., 2004; Walstra, 1993). sion were also evaluated. Furthermore, the theoretical minimum
mean droplet diameter, creaming, the specific surface area and
Dp ¼ 2c=R ð3Þ
the energy dispended to produce the nanoemulsions were
Eq. (3) shows that the amount of energy needed to break the evaluated.
droplets increases when smaller droplets are produced; however, The surfactants used in this study were applied as model surfac-
when lowering the interfacial tension Dp is reduced and therefore tants based on their different charge. Despite, they are commonly
the amount of energy needed to break up a droplet is reduced. used in biotechnology and cosmetics industry, for EFSA, SDS and
Lowering the interfacial tension is one of the roles of the surfac- DTAB cannot be applied in foods, while Tween 20 is consider a food
tants; nevertheless, their most essential role is preventing the coa- additive (EFSA, 2010). Nevertheless, considering FDA regulation
lescence of the newly formed droplets (Schramm, 2006b; Tadros SDS can be applied in food products as surfactant in fruit juice
et al., 2004; Walstra, 1993). drinks under 25 ppm and in coatings on fresh citrus fruit (FDA,
Surfactants preferentially adsorbs to the interfaces, once their 2014a, 2014b). Regarding DTAB, it can be used as an indirect food
molecular structures have non-polar hydrocarbon tails that favor additives: adjuvants, production aids and sanitizers in contact with
non-polar liquids. Lowering the interfacial tension they will mini- food products (FDA, 2014c).
mize the interfacial area between the continuous and dispersed
phases and keep the interfaces smooth (Mason et al., 2006). Low
2. Materials and methods
molecular weight surfactants are able to decrease the interfacial
tension in a greater extension than high molecular weight surfac-
2.1. Materials
tants. This is mainly due differences in the orientation and config-
uration of the surfactants at the interface (Sari et al., 2015). Low
Neobee 1053 medium chain triglycerides (MCTs) is caprylic/-
molecular weight surfactants entirely adsorb and instantaneously
capric triglyceride oil with a fatty acid distribution of 55% of C8:0
orient themselves and the partitioning of the entire molecule
and 44% of C10:0 was kindly provided by Stepan (The
between the two phases facilitates a maximum reduction in the
Netherlands) and was used without further purification. Tween
interfacial tension (Sari et al., 2015). Also, a significant excess of
20 and sodium dodecyl sulphate (SDS) were purchased from
surfactant in the continuous phase is needed this enables the
Sigma–Aldrich (St. Louis, MO, USA) and dodecyltrimethylammo-
new surface area of droplets to be quickly coated during emulsifi-
nium bromide (DTAB) was acquired from Acros Organics (Geel,
cation, inhibiting disruption induced coalescence. This generally
Belgium). Milli-Q water (Milli-Q apparatus, Millipore Corp.,
forms surfactant micelles that dissociate into monomers that
Bedford, MA, USA) was used to prepare all solutions.
rapidly adsorbs to the surface of the droplets (Mason et al., 2006;
Rao and McClements, 2012). After adsorption of the surfactant to
the surface of a droplet, surfactants most provide repulsive forces 2.2. Experimental procedures
strong enough to prevent droplets aggregation. Ionic surfactants
provide a great stability due electrostatic repulsions between dro- 2.2.1. Preparation of non-ionic, cationic and anionic nanoemulsions by
plets. Non-ionic surfactants provide stability due short-range high-pressure homogenization
repulsive forces, such as steric overlap, hydration, thermal fluctua- Oil-in-water (O/W) emulsions were prepared according to
tion interactions, that prevents droplets from getting to close. (Pinheiro et al., 2013) with some modifications. Briefly, the
Briefly, a surfactant must have three characteristics to be effective, nanoemulsions were pre-mixed during 2 min at 5000 rpm using
first, rapidly adsorption to the surface of the new droplets; second, an Ultra-Turrax homogenizer (T 25, Ika-Werke, Germany) fol-
drastically reduce the interfacial tension and third form a mem- lowed by passage through a high-pressure homogenizer (Nano
brane that prevents droplets from aggregating (McClements, DeBEE, BEE International, USA) according to the fractional facto-
2002). rial design (Table 1). To assess the effect of operational condi-
Commercial applications of nanoemulsions are one of the tions on emulsion stability at the nanoscale, samples were
emerging fields of nanotechnology applied to food industry; the evaluated during a 28, 35 and 365 days of storage. The stability
appearance in the food market of nanoemulsion-based food prod- at day 365 was only performed for the nanoemulsions with the
ucts has been growing in the last years. The application of best stability results at day 35. The prepared emulsions were
nanoemulsions in the food industry can be subdivided into five stored at 4 °C in the absence of light, during the evaluation
major categories: fortified beverages, with NutraLease, NovaSol period.

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx 3

Table 1 2.2.5. Nanoemulsion creaming rate

Independent variables used in the 24 fractional factorial design: pressure of Ten grams of nanoemulsion were transferred into a 15 mL
homogenization (expressed in terms of Psi), number of cycles between each homog-
enization, surfactant concentration (expressed in % wt) and oil to water (O/W) volume
conical centrifuge tube (Falcon™, Fisher Scientific), tightly sealed
ratio (expressed in % vol). with a plastic cap, and then stored at 4 °C) for approximately
24 h, 1 month, 3 months and 1 year. After storage, several emul-
Sample Pressure (Psi ⁄ 1000) Cycles Surfactant O/W ratio (% vol)
sions separated into an opaque layer at the top, a turbid layer in
1 10 10 0.5 5 the middle, and a transparent layer at the bottom. The ‘‘serum
2 20 10 0.5 15
3 10 30 0.5 15
layer’’ was defined to be the sum of the turbid and transparent
4 20 30 0.5 5 layers. The total height of the emulsion (HE) and the height of
5 10 10 1.5 15 the serum layer (HS) were measured using a ruler (Li et al.,
6 20 10 1.5 5 2010).
7 10 30 1.5 5
8 20 30 1.5 15
9 15 20 1 10 2.2.6. Microscopy
10 15 20 1 10
The morphology of nanoemulsions was evaluated by transmis-
11 15 20 1 10
sion electron microscopy (TEM) (EM 902A, ZEISS, Germany) oper-
ating at 80 kV. TEM samples were prepared by depositing the
nanoemulsion suspensions on a carbon-coated copper grid, and
2.2.2. Kinetic stability studies
negatively stained with 1% (w/v) uranyl acetate for observation.
Kinetic stability studies were performed through centrifugation,
Samples were air-dried before analyses.
heating–cooling cycle and thermal stress as followed:
Centrifugation: formulations were centrifuged at 5000g for
30 min. Those formulations that did not show any phase separa- 2.2.7. Nanoemulsions viscosity measurements
tion were taken for the heating–cooling cycle test. Viscosity measurements of the nanoemulsion samples were
Heating–cooling cycle: formulations were subjected to six cycles conducted using a dynamic shear rheometer (TA Instruments,
between 4 °C and 45 °C with a storage time higher than 48 h at New Castle, DE) with a shear rate profile from 0.1 to 1000 s1. All
each temperature. Those formulations, which were stable at these the measurements were performed at 25 °C. The values of these
temperatures cycles, were further taken for size measurements measurements were applied to Eqs. (7) and (8).
(Shafiq et al., 2007).
Thermal stress: The influence of temperature on nanoemulsions
2.2.8. Nanoemulsions density measurements
stability was examined through Dynamic Light Scattering (DLS)
Density measurements of the nanoemulsion samples were con-
technique, by measuring the Hd of the nanoemulsions. Briefly,
ducted using a density-meter Densito 30PX (Mettler-Toledo, Inc.,
1 mL of the formulations were submitted to a range of tempera-
Columbus, OH) at 25 °C. The values of these measurements were
tures from 20 to 80 °C, with increasing intervals of 10 °C, during
applied to Eqs. (6)–(8).
10 min at each condition (Morais Diane and Burgess, 2014). The
temperature was adjusted using a Nano ZS-90 equipment
(Zetasizer Nano ZS-90, Malvern Instruments, Worcestershire, UK). 2.2.9. Nanoemulsions interfacial tension measurements
The data was reported as the mean droplet diameter (hydrody- The interfacial tension of the nanoemulsions solution was mea-
namic diameter, Hd), mean ± SD (n = 3). sured by the pendant drop method using the Laplace–Young
approximation, (Song and Springer, 1996), with a face contact
angle meter (OCA 20, Dataphysics, Germany). The samples of the
2.2.3. Nanoemulsion size measurements coatings were taken with a 500 lL syringe (Hamilton,
The particle size distribution and Polydispersity Index (PdI) of Switzerland), with a needle of 0.75 mm of diameter. Three repli-
nanoemulsions were determined using DLS (Zetasizer Nano cates were performed at 20 °C. The values of these measurements
ZS-90, Malvern Instruments, Worcestershire, UK). The nanoemul- were applied to Eqs. (1)–(3).
sion samples were diluted 10 in distilled water at ambient
temperature. The data was reported as the mean droplet diame-
2.3. Statistical procedures
ter (hydrodynamic diameter, Hd). Polydispersity Index is a
dimensionless and indicates the heterogeneity (monodisperse
2.3.1. Experimental design
or polydisperse) of sizes of particles in a mixture (Malvern,
A multifactor model (24 fractional factorial design) was used to
2011). Each sample was analyzed in a disposable polystyrene
evaluate the independent variables (see Table 1): homogenization
cell (DTS0012, Malvern Instruments). The measurements were
pressure (10,000–20,000 Psi), surfactant concentration (0.5–1.5%
performed in duplicate, with three readings for each of them.
w/w), ratio between oily and water phase (O/W ratio) (5–15%
The results are given as the average ± standard deviation of the
v/v) and cycles between each homogenization (10–30) and evalu-
six values obtained (Rao and McClements, 2013; Silva et al.,
ate the main effects and interaction effects of these formulations
2011).
on the hydrodynamic diameter, polydispersity index and zeta
potential. A total of 11 experiments for each surfactant were con-
ducted with 3 center points (in order to allow the estimation of
2.2.4. Nanoemulsion charge measurements
pure error).
The droplet charge (zeta potential) of the nanoemulsions was
determined using a particle micro-electrophoresis instrument
(Zetasizer Nano ZS-90, Malvern Instruments, Worcestershire, UK). 2.3.2. Data analyses
Samples were diluted 100 in distilled water prior to measure- Data analyses were performed using Microsoft Windows Excel
ments in order to avoid multiple scattering effects at ambient tem- 2011, using the Tukey’s Multiple Comparison Test with a confi-
perature and placing the diluted emulsions into disposable dence interval of 95% in GraphPad Prism 5 (GraphPad Software,
capillary cells (DTS 1060, Malvern Instruments) (Ozturk et al., Inc.) and using ANOVA in STATISTICA 7.0 (Statsoft, Tulsa, OK,
2014; Rao and McClements, 2013). USA).

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
4 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx

3. Results McClements, 2011). Also, lowering the interfacial tension through

the use of surfactants may facilitate droplet deformation and dis-
3.1. Effect of process conditions on nanoemulsions ruption (Brösel and Schubert, 1999).
As shown in Table 2 the mean Hd of nanoemulsions decreases
The combination of surfactant and a high-energy homogeniza- with the increase of homogenization pressure and number of
tion process (using an Ultra-Turrax as a premix step followed by cycles. This behavior is explained by the high intensity of shear
a high-pressure homogenization process) allowed the formation forces produced during the homogenization process (Tan and
of nanoemulsions with different characteristics. For an initial Nakajima, 2005). Also, increasing the homogenization pressure
screening, the hydrodynamic diameter (Hd), the polydispersity and the number of cycles will lead to a higher specific energy, Ev
index (PdI) and the zeta potential (Zp) of nanoemulsions are the (Eq. (4)), delivered to the emulsion by the high-pressure homoge-
most studied parameters, once they give an accurate idea about nizer, which is the mechanical energy input per unit of volume of
the formed nanoemulsions and their stability (Cheong et al., the chamber where the droplets are disrupted. Ev can be calculated
2008; Sood et al., 2014; Tan and Nakajima, 2005). by the homogenization pressure (DP) times the number of cycles
Table 2 shows the values of Hd for the nanoemulsions produced (n) (Donsì et al., 2011; Stang et al., 2001). This indicates that fur-
with Tween 20, SDS and DTAB. Values ranged between 171 and ther increases of the dispersion energy levels lead to even smaller
341 nm for nanoemulsions produced with Tween 20; between nanoemulsions (Donsì et al., 2011; Silva et al., 2011).
126 and 177 nm when produced with SDS and between 135 and
198 nm for DTAB. Transmission Electron Microscopy (TEM) con-
Ev ¼ DP  n ð4Þ
firmed the development of these nanoemulsions and validated
the mean droplet diameters achieved; Fig. 1 represents a TEM Tables 2 and 3 show that the increase of surfactant concentra-
microphotograph of the anionic, cationic and non-ionic nanoemul- tions leads to lower values of Hd. This is due to the fact that more
sions produced using the high-pressure homogenization technique surfactant is present to cover any new droplet surfaces formed dur-
at 1% wt of surfactant, dispersed in the aqueous phase, and MCTs ing homogenization and because the droplet surfaces will be cov-
oil as the organic phase at 15,000 Psi, during 20 cycles. Also, TEM ered more quickly by the surfactants (Qian and McClements,
was performed to observe the morphology of the different charge 2011). Also, higher surfactant concentrations increase the interfa-
nanoemulsions. In Fig. 1a it is possible to see oil-droplet aggrega- cial area, while reducing the interfacial tension, leading to lower
tions due dehydration conditions after the use of uracyl. droplet sizes (Anton et al., 2007; Esquena et al., 2003; Liu et al.,
In order to study the most influent factors on the dependent 2006).
parameters (Hd, PdI and Zp) a multifactor model 24 was used, being Table 2 shows the PdI values obtained for nanoemulsions, rang-
the results analyzed using ANOVA, which identifies the significant ing from 0.164 to 0.242, 0.119 to 0.179 and 0.146 to 0.202 for
factors that affect the responses (Table 3). Results show that Hd is Tween 20, SDS and DTAB, respectively. The efficiency of droplet
significantly influenced (p < 0.05) by the homogenization pressure disruption increased for higher specific energy (i.e. homogeniza-
when SDS is used as surfactant; nevertheless, when DTAB was used tion pressure times the number of cycles), at the most severe con-
as surfactants for the nanoemulsion production, Hd is significantly ditions the smallest mean droplet diameter was achieved,
influenced (p < 0.05) by homogenization pressure, surfactant con- nevertheless for the PdI this did not happen. This is due the fact
centration and O/W ratio. In the case of nanoemulsions produced that PdI may also depend on emulsifier surface and dynamic prop-
with Tween 20, Hd is significantly influenced (p < 0.05) by homog- erties (Donsì et al., 2011). Donsì et al. (2011) also verified that PdI
enization pressure, the interaction between the homogenization was mainly affected by the kinetics of surface adsorption of the
pressure and the O/W ratio, by the surfactant concentration and emulsifier.
the number of cycles. The nanoemulsions produced with SDS ANOVA results show that the homogenization pressure signifi-
achieved the smallest value of Hd, followed by the nanoemulsions cantly influenced (p < 0.05) the PdI values of nanoemulsions pro-
produced with DTAB and Tween 20, respectively. The differences in duced with the different surfactants (SDS, DTAB and Tween 20),
Hd values for the three surfactants used can be attributed to the as shown in Table 3. For the nanoemulsions produced with
molecular weight (Mw) of the surfactants (Mw SDS < Mw Tween 20 the interaction between pressure and O/W ratio and
DTAB < Mw Tween 20), in general small molecule surfactants can the interaction between pressure and the number of cycles also
lead, under similar processing conditions, to smaller Hd values, significantly influenced (p < 0.05) PdI values. Regarding the
which has been attributed to differences in adsorption rates (faster nanoemulsions produced with DTAB, the interaction between
adsorption rates result in smaller sizes) and in interfacial proper- pressure and O/W ratio and the interaction between pressure
ties (such as thickness, charge, permeability, and environmental and the number of cycles also significantly influenced (p < 0.05)
responsiveness) (Donsì et al., 2011; McClements, 2010; Qian and PdI, as also did the number of cycles (p < 0.05).

Table 2
Experimental Hd, PdI and Zp values for nanoemulsions produced with SDS, Tween 20 and DTAB as surfactants immediately after production, for the fractional factorial design.

Sample SDS day 0 DTAB day 0 Tween 20 day 0

Hd (nm) Zp (mV) PdI Hd (nm) Zp (mV) PdI Hd (nm) Zp (mV) PdI
1 168.6 ± 1.1 69.4 ± 1.2 0.135 ± 0.01 174.9 ± 5.0 64.8 ± 0.3 0.158 ± 0.01 257.8 ± 2.7 15.3 ± 1.0 0.242 ± 0.01
2 141 ± 4.6 66.2 ± 1.8 0.150 ± 0.02 197.6 ± 5.5 43.0 ± 0.5 0.195 ± 0.02 171.2 ± 4.6 30.1 ± 4.8 0.164 ± 0.01
3 147.7 ± 3.5 69.3 ± 1.2 0.131 ± 0.01 187 ± 5.6 43.1 ± 2.4 0.202 ± 0.01 341.2 ± 8.1 23.1 ± 0.7 0.224 ± 0.02
4 128.9 ± 3.7 78.7 ± 1.7 0.153 ± 0.01 172.4 ± 4.8 71.0 ± 1.2 0.147 ± 0.01 323.3 ± 4.7 16.6 ± 3.5 0.210 ± 0.01
5 177.1 ± 3.9 73.1 ± 2.3 0.179 ± 0.02 163.0 ± 4.7 71.2 ± 2.5 0.148 ± 0.01 277.9 ± 8.8 25.0 ± 0.5 0.236 ± 0.01
6 136.5 ± 5.3 68.2 ± 1.9 0.130 ± 0.01 141.2 ± 3.4 70.3 ± 2.9 0.146 ± 0.01 256.2 ± 10.5 20.1 ± 2.4 0.181 ± 0.02
7 157.5 ± 3.6 84.2 ± 0.4 0.119 ± 0.01 134.9 ± 2.9 73.1 ± 3.1 0.171 ± 0.01 215.5 ± 6.1 27.5 ± 1.0 0.177 ± 0.03
8 126.6 ± 3.8 73.7 ± 2.2 0.148 ± 0.02 153.9 ± 4.8 74.6 ± 2.2 0.179 ± 0.01 195.8 ± 5.8 20.4 ± 2.2 0.197 ± 0.01
CP 140.6 ± 12.3 73.5 ± 3.8 0.145 ± 0.02 156.3 ± 2.7 75.7 ± 0.5 0.197 ± 0.02 237.4 ± 6.4 24.7 ± 2.0 0.192 ± 0.01

Each value represents mean ± SD (n = 3); CP – central point; Hd – hydrodynamic diameter; Zp – zeta potential; PdI – polydispersity index.

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx 5

Fig. 1. TEM microphotograph of negatively stained nanoemulsions with uracyl 1% w/w. (a) Anionic nanoemulsion; (b) cationic nanoemulsion and (c) non-ionic
nanoemulsion.

Table 3 non-ionic, anionic and cationic emulsifiers, respectively (Pinheiro

ANOVA results for dependent parameters estimation. et al., 2013).
Independent variable p-value p-value p-value
SDS Tween DTAB 3.2. Kinetic stability of nanoemulsions during storage
Hd Pressure 0.047301 0.014975 0.028920
Pressure by O/W ratio – 0.002528 – It is important to distinguish thermodynamic stability from
Surfactant – 0.014530 0.004790 kinetic stability; while thermodynamics will tell if a process will
concentration
or not occur (no matter the time to complete it), kinetics will tell
Cycles – 0.024731 –
Ratio – – 0.013720 the rate (what time scale) and the degree of change, if it occurs.
Kinetic stability tests were carried out to eliminate any unstable
PdI Pressure 0.048942 0.013272 0.030854
Pressure by O/W ratio – 0.040542 0.010941 formulation, due to creaming or sedimentation. Tests were per-
Pressure by cycles – 0.011017 0.022267 formed in terms of centrifugation and through heating and cooling
Cycles – – 0.009264 cycles. Only those formulations that were found to be stable were
Zp Pressure 0.041880 – – selected for evaluation of Hd, PdI and Zp during storage (Table 4).
Pressure by cycles – 0.049921 –
3.2.1. Size distribution
Regarding centrifugation tests, all produced formulations (using
The influence of processing conditions on Zp was also evaluated, all type of surfactants) showed no macroscopic instability phe-
being of extremely importance due to the different charges of the nomena (i.e. creaming or phase separation), and the sizes remained
surfactants used. Zp values ranged from 15 to 30 mV for unchanged after centrifugation (data not shown), due to this all the
nanoemulsions produced with Tween 20, a non-ionic surfactant. formulations were subjected to the heating–cooling cycles. Once
The high values of Zp obtained for the nanoemulsions produced more the produced nanoemulsions showed no macroscopic insta-
with Tween 20 may be explained by the composition of the bility phenomena (i.e. creaming or phase separation) and the sizes
MCTs oil (free fatty acids and salts) and by the presence of MCTs remained unchanged (data not shown) after the heating–cooling
oil at the interface of the surfactant, even at equilibrium, there is cycles’ tests. Therefore, all formulations were evaluated by means
a continuous exchange of molecules between the interior of the of size stability during 28 days of storage, in order to evaluate
droplets, the continuous phase and the interface, that will occur the influence of the process conditions in the mean droplet diam-
at a rate dependent of the mass transport of the molecules through eter during storage.
the system (McClements, 2005). For nanoemulsions produced with Table 4 shows the values of Hd, PdI and Zp of nanoemulsions
SDS, an anionic surfactant, Zp values ranged between 66 and after 28 days of storage. In general, it can be observed that these
84 mV. The Zp of the nanoemulsion has a negative value due to parameters remain unchanged after 28 days of storage. For the
the presence of head sulfate groups of SDS molecules (Jadhav nanoemulsions produced with the anionic surfactant, SDS, only
et al., 2015). Nanoemulsions produced with DTAB (cationic surfac- sample 5 shows an increase in size of 13 nm. Donsì et al. (2011)
tant) displayed Zp values ranging from 43 to 76 mV. ANOVA results showed that SDS as a significantly short characteristic time for
showed that the process parameters did not significantly influ- adsorbing to the oil-in-water interface, when compared to surfac-
enced Zp for all the developed nanoemulsions (Table 3). These elec- tants like polysorbates (e.g. Tween 20) (Donsì et al., 2011). Fast
trical charges were expected as Tween 20, SDS and DTAB are adsorption kinetics of the surfactant to the interface can

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
6 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx

Table 4
Experimental Hd values obtained for nanoemulsions produced with SDS, Tween 20 and DTAB as surfactants after 28 days of storage for the fractional factorial design.

Sample SDS day 28 DTAB day 28 Tween 20 day 28

Hd (nm) Zp (mV) PdI Hd (nm) Zp (mV) PdI Hd (nm) Zp (mV) PdI
1 159.8 ± 4.4 61.2 ± 1.1 0.136 ± 0.03 172.0 ± 5.1 63.1 ± 0.8 0.168 ± 0.01 174.2 ± 2.4 14.7 ± 0.1 0.194 ± 0.01
2 137.8 ± 3.0 62.2 ± 1.3 0.141 ± 0.01 175.7 ± 1.4 58.3 ± 0.6 0.174 ± 0.01 166.1 ± 4.5 32.3 ± 0.3 0.134 ± 0.01
3 147.4 ± 3.8 62.3 ± 1.9 0.121 ± 0.01 194.8 ± 4.2 60.7 ± 0.9 0.185 ± 0.02 195.5 ± 6.8 22.8 ± 1.1 0.197 ± 0.01
4 130.9 ± 4.6 94.4 ± 1.9 0.165 ± 0.01 168.2 ± 2.0 67.0 ± 0.7 0.162 ± 0.01 253.9 ± 7.4 15.1 ± 0.7 0.213 ± 0.01
5 164.6 ± 6.9 63.5 ± 1.7 0.168 ± 0.03 159.9 ± 4.8 66.3 ± 1.9 0.156 ± 0.01 203.6 ± 3.7 27.8 ± 0.4 0.131 ± 0.02
6 132.58 ± 2.6 76.7 ± 2.3 0.141 ± 0.02 118.9 ± 1.5 76.8 ± 1.5 0.155 ± 0.02 245.9 ± 5.2 19.5 ± 0.2 0.173 ± 0.01
7 159.4 ± 6.5 62.5 ± 1.0 0.138 ± 0.01 142.9 ± 4.4 70.8 ± 1.9 0.227 ± 0.03 177.8 ± 4.2 24.3 ± 0.3 0.140 ± 0.01
8 126.1 ± 5.2 92.6 ± 1.8 0.139 ± 0.01 152.0 ± 3.1 74.9 ± 1.0 0.187 ± 0.01 196.7 ± 0.2 25.7 ± 2.3 0.186 ± 0.01
CP 140.5 ± 4.0 78.2 ± 1.4 0.155 ± 0.01 155.4 ± 3.0 73.6 ± 1.9 0.196 ± 0.01 204.6 ± 3.7 26.3 ± 1.3 0.125 ± 0.01

Each value represents mean ± SD (n = 3); CP – central point; Hd – hydrodynamic diameter; Zp – zeta potential; PdI – polydispersity index.

significantly reduce the re-coalescence phenomena. This conju- storage. The surfactant concentration and the oil content did not
gated with the lower oil-in-water interfacial tension lead to stable show statistically significant differences in terms of Hd.
nanoemulsions with size values ranged between 130 and 165 nm, For the cationic nanoemulsions the most influent parameters
depending on the processing conditions (Donsì et al., 2011). were surfactant concentration (p = 0.0048), followed by ratio
The same behavior was observed for nanoemulsions produced between oily and aqueous phases (p = 0.014) and pressure
with DTAB (cationic surfactant), where the size values remained (p = 0.0289). The increase of surfactant concentration leads to the
unchangeable during storage time. Li and McClements (2011) pre- formation of smaller nanoemulsions, once surfactant adsorption
pared stable nanoemulsions using DTAB, achieving size values of to the interface may lower the interfacial tension, resulting in a
approximately 187 nm, indicating that these nanoemulsions were higher mechanical resistance to rupture (McClements, 2005).
not susceptible to droplet aggregation; those authors also claim Low oil content leads to the decrease of droplet size; this can be
that the presence of DTAB in low concentrations increased the rate explained by the influence of oil in nanoemulsions viscosity (i.e.
and extent the lipid digestion. One possible explanation is the fact low oil content leads to lower viscosity values). Higher viscosities
that this surfactant may facilitate the adsorption of lipase to the induce a flow resistance in the chamber of the high-pressure
nanoemulsion surface due to electrostatic attraction (Li and homogenizer diminishing the rate and efficiency of droplet disrup-
McClements, 2011). For nanoemulsions produced with the tion, leading to higher mean droplet sizes (McClements, 2005;
non-ionic surfactant, Tween 20, an increase in the size values of Troncoso et al., 2012). For the nanoemulsions produced with
samples 1, 3, 4, 5 and 7 after storage was observed. This increase non-ionic surfactant the significant parameters were the combina-
could be explained by the breakdown process of the emulsions, tion of pressure with oil content (p = 0.0025), the surfactant con-
namely the creaming effect, where the larger droplets move faster centration (p = 0.0145), pressure (p = 0.0150) and the number of
to the top of the container, due the fact that when the gravitational cycles (p = 0.0247). Here the increase in oil content lead to smaller
forces exceed the thermal motion of the droplets (Brownian Hd values, contrary to the usual assumption that the increase of the
motion) a concentration gradient is built up (Tadros, 2013). Silva oil content increases Hd values due to the influence of oil in the vis-
et al. (2011) also found a similar behavior for nanoemulsions pro- cosity (McClements, 2005; Troncoso et al., 2012). An increase in oil
duced with Tween 20. They showed that the samples with lower content results in the increase of the concentration of droplets (i.e.
size values had a greater tendency to aggregate, since they were increase of the dispersed phase volume fraction) in an emulsion,
more susceptible to Brownian motion, which leads to a superior which leads to higher frequency of droplet collisions and hence
chance of collision, allowing aggregation to become a dominant coalescence during emulsification. Also, the viscosity of the emul-
mechanism for emulsion instability (Silva et al., 2011). The larger sion increases, which may change the flow from turbulent to lam-
droplets produced when Tween 20 is used as surfactant can lead inar. The presence of more particles results in an increase of the
to lower nanoemulsion stability as measured in terms of particle velocity gradient. In turbulent flow, the increase of oil content
size. These larger droplets result in lower specific surface areas, can induce turbulence depression leading to larger droplet sizes.
being more susceptible to surface area fluctuations, increasing If the ratio between the surfactant and the continuous phase is
creaming and coalescence phenomena (Tadros, 2013). constant, an increase in oil content results in a decrease of the sur-
factant concentration, resulting in larger droplets. Nevertheless, if
3.3. Selection of the most suitable process conditions the ratio between the surfactant concentration and the disperse
phase is kept constant, then this is reversed, so at this point it is
The results showed that for anionic nanoemulsions the smallest impossible to draw any conclusions regarding the mechanisms
sizes were achieved using the highest pressure and highest num- that may come into play (McClements, 2005; Schramm, 2006a;
ber of cycles tested, 20,000 Psi and 30 cycles of homogenization Tadros, 2013).
(Table 2). In fact, a higher homogenization pressure and number The selection of the most suitable process conditions for the
of cycles increase the specific energy delivered by the three surfactants was rather difficult since for each surfactant dif-
high-pressure homogenizer, which is responsible for increasing ferent process parameters had different responses. However, due
the disruption rates that, in the presence of enough surfactant, will to the need to select one formulation for further study the central
lead to smaller sizes (Donsì et al., 2011). Nevertheless, performing point formulation was selected (15,000 Psi, 20 cycles, 1% wt of sur-
a one-way ANOVA to the obtained size values it was possible to factant and an oil content of 10% wt). The pressure of 20,000 Psi
verify that there were no statistically significant differences and the 30 cycles of homogenization were not considered, due to
between the results achieved with 15,000 and 20,000 Psi and the overheating that these conditions induced to the nanoemulsion
between 20 and 30 cycles (p = 0.2269). Being so, it can be con- solution, that could be harmful for bioactive compounds that are
cluded that 15,000 Psi and 20 cycles are the most suitable process heat sensitive. Also, increasing the pressure and the number of
conditions to carry on the study of creaming during long-term cycles increases the energy consumption, which can increase the

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx 7

price of producing nanoemulsions. An oily/aqueous phase volume

ratio of 10% wt was used according to Silva et al. (2011) and Tan
and Nakajima (2005). Tan and Nakajima (2005) studied the influ-
ence of phase ratios (1:9 and 2:8); as the oil content increases,
the available surfactant decreases, reducing the number of mole-
cules able to stabilize the formed droplets, thus favoring coales-
cence and resulting in higher mean droplet diameters. In
addition, higher viscosity of the oily phase leads to a less favorable
mixing efficiency (Tan and Nakajima, 2005).

3.4. Kinetic stability for the selected conditions

Kinetic stability tests are useful not only to provide information

about the short-term stability of nanoemulsions, but also for pre-
dicting the long-term stability. The input of high temperature
and centrifugal forces increases Brownian motion and allows the Fig. 2. Hydrodynamic diameter (Hd) and polydispersity index (PdI) during 35 days
dispersed droplets to approach one another (McClements, 2005; of storage. Bars indicate standard deviation (n = 3). Lines are for readers’ guidance
Morais Diane and Burgess, 2014; Schramm, 2006b). Information’s and do not represent a model prediction.
such as the rate at which the properties of a nanoemulsion change
with time (i.e. time dependence) is highly important for food
industry (McClements, 2005; Schramm, 2006b; Walstra, 1993). Here, C is the surface load of the surfactant at saturation
The selected nanoemulsion formulations were studied in terms (kg/m2), / is the disperse phase volume (dimensionless), and C0 s
of centrifugation, heating–cooling cycles, creaming and size stabil- is the concentration of surfactant at the continuous phase
ity during storage (35 days) at 4 °C. Creaming and size stability (kg/m3) (Troncoso et al., 2012). The surface load of SDS, DTAB
during storage were also evaluated after one year of storage at and Tween 20 were assumed as 1, 1 and 1.5 mg/m2, respectively
4 °C, in the absence of light. (Berton-Carabin et al., 2014). The minimum theoretical droplet
Nanoemulsions produced with anionic, cationic and non-ionic diameter achieved was 67 nm for the anionic and cationic
surfactants showed no macroscopic sign of instability phenomena nanoemulsions and 100 nm for the non-ionic nanoemulsion. The
(i.e. creaming or phase separation) after centrifugation. The results experimental mean droplet diameter obtained in our study was
showed that the nanoemulsions produced with the anionic, catio- 137 nm for the anionic nanoemulsion, 152 nm for the cationic
nic and non-ionic surfactants maintained their mean droplet size nanoemulsion and 223 nm for the non-ionic nanoemulsion.
and hence were stable even under a centrifugal stress of 5000g These results are far from the theoretical ones, which suggests that
(Table 5). Results showed that these nanoemulsions under cen- there was not enough surfactant initially present to cover all dro-
trifugal tests did not present statistically significant (p > 0.05) dif- plets, or the homogenizer was incapable of generating enough dis-
ferences in the values of Hd when compared to those maintained ruptive forces for the process conditions used (Qian and
at 4 °C (see Table 5). Similar results were obtained following the McClements, 2011; Troncoso et al., 2012). The oil viscosity has
heating–cooling tests (Table 5). an important role in the mean droplet sizes. Troncoso et al.
For the same process and formulation conditions nanoemul- (2012) showed that for more viscous oily phases the
sions produced with SDS achieved lower sizes compared to the high-pressure homogenizer used to prepare the nanoemulsions
cationic or non-ionic surfactant nanoemulsions (Fig. 2). The differ- was incapable of generating sufficiently intensive disruptive forces,
ences in Hd values for the different surfactants can be attributed to while reducing the viscosity of the oil lead to similar results
the molecular weight (Mw) of the surfactants (Mw SDS < Mw between the experimental work and the theoretical data
DTAB < Mw Tween 20). In general small molecule surfactants can (Troncoso et al., 2012).
lead, under similar processing conditions, to smaller Hd values Also the ability to lower oil:water interfacial tension and the
(Donsì et al., 2011; Qian and McClements, 2011). Within the faster kinetics of adsorption at the oil:water interface can explain
homogenizer chambers small molecule surfactants are able to the differences in the mean droplet size (Donsì et al., 2011). A
quickly adsorb to the newly formed droplets surface (Qian and decrease in interfacial tension, while leading to lower droplet sizes,
McClements, 2011; Troncoso et al., 2012). also increases the surface area of nanoemulsions. The properties of
The theoretical minimum size of the produced nanoemulsions colloidal systems’ surface are very important and lead to unique
through high-pressure homogenization was calculated using Eq. physical properties (Schramm, 2006a, 2006b, 2006c). The specific
(5). surface area of nanoemulsions, considering n spheres of density
q and radius R was determined (Eq. (6)) for the selected conditions
Rmin ¼ ð3  C  /Þ=ðcs ð1  /ÞÞ ð5Þ (Schramm, 2006a).

Table 5
Experimental Hd and PdI values for nanoemulsions produced with SDS, Tween 20 and DTAB as surfactants after kinetic stability tests.

Treatment SDS DTAB Tween 20

Hd (nm) PdI Hd (nm) PdI Hd (nm) PdI
a a a a a
After production 137.4 ± 0.6 0.128 ± 0.04 152.1 ± 2.1 0.178 ± 0.02 223.2 ± 1.4 0.218 ± 0.01a
After centrifugation 134.8 ± 2.4a 0.126 ± 0.03a 155.4 ± 2.1a 0.177 ± 0.01a 225.6 ± 1.1a 0.221 ± 0.02a
After heating–cooling cycles 136.3 ± 3.0a 0.121 ± 0.03a 155.7 ± 2.6a 0.176 ± 0.01a 221.2 ± 4.6a 0.217 ± 0.01a
After 1 year of storage 138.5 ± 1.3a 0.129 ± 0.05a n.a. n.a. n.a. n.a.

Each value represents mean ± SD (n = 3). Tukey’s Multiple Comparison Test with a confidence interval of 95% was applied; n.a. – not available; Hd – hydrodynamic diameter;
Zp – zeta potential; PdI – polydispersity index.
a
Different letters between rows means statistically different.

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
8 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx

Asp ¼ 3=qR ð6Þ motion can exceed the gravity force and therefore creaming is pre-
vented (Tadros, 2013). The smaller size obtained by the anionic
The combination of high-pressure homogenization with the
nanoemulsions is one of the reasons that allow explaining why
intrinsic properties of the anionic surfactant lead to a specific sur-
these nanoemulsions had higher stability than the cationic and
face area of 44.3 m2/g of oil promoting, promoting an increase of
non-ionic nanoemulsions. Also, increasing the surface charge of
ca. 37,000 times of the surface area which was of 0.0012 m2/g of
the droplets can slow down the creaming rate because of the
oil before the high-pressure homogenization process. Higher speci-
increasing repulsive forces between the droplets (McClements,
fic surface areas (smaller droplets) are less susceptible to surface
2005).
fluctuations and hence creaming and coalescence are reduced,
After 1 year of storage at 4 °C, the nanoemulsions produced
explaining the high stability of nanoemulsions (Tadros, 2013). As
with cationic and non-ionic surfactants showed visible creaming,
expected, due to the higher droplet sizes, cationic and non-ionic
presenting a creaming height of 0.25 and 0.5 cm (data not shown),
surfactants nanoemulsions had lower specific surface areas. The
respectively. The nanoemulsion produced with the anionic surfac-
cationic nanoemulsion had 1.1 times less surface area when com-
tant remained kinetically stable, without evidence of creaming
pared to the anionic nanoemulsion, achieving a specific surface
and/or phase separation, showing an Hd of 138.5 nm.
area of 39.7 m2/g of oil. Non-ionic nanoemulsion had 1.9 times less
surface area when compared to the anionic nanoemulsion, achiev-
ing a specific surface area of 22.6 m2/g of oil. 3.5. Effect of temperature in size stability
Consumers expect homogenous products, without the undesir-
able separation of emulsion into oily (opaque and rich in droplets) The thermal stability tests performed to select nanoemulsions
and aqueous phases (less opaque and without droplets). In order to (please see topic 3.3.) showed no macroscopic sign of instability
evaluate emulsion homogeneity, creaming was determined after phenomena (i.e. creaming or phase separation) after heating the
1 year of storage at 4 °C. Using Stokes equation, where the terminal samples from 20 to 80 °C. For this test, Hd measurements were per-
settling velocity (dx/dt) is proportional to gravity (g) and the formed immediately after nanoemulsions’ preparation and follow-
square of particle size (a), and inversely proportional to the fluid ing the 30 min of heating at each temperature (20–80 °C, refer to
viscosity (g), it is possible to predict the rate of creaming once it Table 6). Temperature increase did not have an immediate effect
is proportional to the settling velocity. in the characteristics of nanoemulsions, e.g. statistical analysis
showed that thermal stress did not provoke significant differences
ðdx=dtÞ ¼ ð2a2 ðq2  q1 ÞgÞ=ð9gÞ ð7Þ in the mean hydrodynamic diameter (Table 6). This is due the fact
that unlike microemulsions, that change their morphology, size
Stokes law assumes that emulsions are uncharged and spheri-
and shape as a function of temperature and/or composition,
cal; nevertheless, since the droplets will interact in the case of
nanoemulsions stability is less sensitive to temperature increase,
charged nanoemulsions stabilized with surfactants or polymers
and this change does not promote a modification in the continuous
(opposite to what is assumed in Stokes’ theory), underestimations
phase, being the droplet sizes unchangeable (Anton and
or overestimations of the terminal velocity are to be expected
Vandamme, 2011; Gordon et al., 2014).
(Schramm, 2006a). When the emulsions are charged an electrical
potential is created; this speeds up the counter-ions and slows
down the droplets, although this effect can be quickly dissipated 3.6. Energy consumption
at high enough electrolyte concentrations. Applying a correction
factor to the Stokes’ law it is possible to determine the settling One of the drawbacks of high-energy methods like
velocity. Richardson–Zaki equation (Eq. (8)) was applied for this, high-pressure homogenization is the significant amount of energy
where U is the dispersed-phase volume fraction (U < 1), for parti- required during the process (Tadros et al., 2004; Walstra, 1993). It
cles smaller than 1 lm in diameter the exponent found was 5.25 is therefore important to evaluate the interfacial energy (Eq. (1)),
(Barnes and Holbrook, 1993; Schramm, 2006a). the mechanical energy (Eq. (4)) and the % of energy dissipated into
heat. Results showed that mechanical energy exceeded interfacial
0
ðdx=dtÞ ¼ ðdx=dtÞ  ð1  /Þ5:25 ð8Þ energy by several orders of magnitude. The interfacial energy val-
ues required for the formation of anionic, cationic and non-ionic
Anionic nanoemulsions showed no macroscopic sign of cream-
nanoemulsions were 1.97, 1.92 and 0.91 J, respectively, while the
ing or phase separation (Table 5); nevertheless, cationic and
mechanical energy was 2070 MJ for all the nanoemulsions. This
non-ionic nanoemulsions presented a phase separation due to
means an efficiency of 0.1%, while 99.9% of the energy was dissi-
creaming with a height of 0.25 and 0.5 cm after 1 year of storage,
pated as heat. These results are in agreement with the results pre-
respectively. The obtained creaming values for the nanoemulsions
sent by Tadros et al. (2004) and Walstra (1993).
produced with Tween 20 are in agreement with the predicted val-
ues (0.42 cm/year); nevertheless, for anionic and cationic
nanoemulsions the same behavior was not verified. The predicted 4. Conclusions
value for the cationic nanoemulsion was 0.09 cm/year, opposing to
the obtained value of 0.25 cm/year. One possible explanation for Homogenization pressure, surfactant type and oil concentration
this discrepancy is that even reducing droplet sizes to 155 nm this were found to be critical to achieve the desired hydrodynamic
was not enough to overcome the density difference between dis- diameters, polydispersity index and zeta potential of nanoemul-
persed and continuous phases. Here the gravitational forces exceed sions. Although nanoemulsions were produced at all levels of pro-
the Brownian motion, creating a concentration gradient that makes cess and formulation parameters being the parameters’ values for
the larger droplets quickly move to the top of the container the preparation of very stable nanoemulsions with the smallest
(Schramm, 2006a; Tadros, 2013). For the anionic nanoemulsion hydrodynamic diameter: homogenization pressure of 20,000 Psi,
the predicted creaming value was 0.07 cm/year; nevertheless, no 20 cycles of homogenization, 1% wt of surfactant and 10% wt of
macroscopic sign of instability phenomena (i.e. creaming or phase oil. Results showed that differently charged nanoemulsions were
separation) was visible. As the gravity force is proportional to R3, if kinetically stable during storage, 28 days (for fractional factorial
R is reduced 10 times the gravity force is reduced by 1000 times. design) and 35 days (for the selected formulations) at 4 °C in terms
Below 100 to 200 nm of droplet size (which also depends on the of hydrodynamic diameter, polydispersity index and zeta potential.
density difference between the oily and water phases), Brownian However, the anionic nanoemulsion displayed the highest stability

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx 9

Table 6
Experimental Hd and PdI values for nanoemulsions produced with SDS, Tween 20 and DTAB as surfactants after thermal stress tests.

Treatment SDS DTAB Tween 20

Hd (nm) PdI Hd (nm) PdI Hd (nm) PdI
After production 137.5 ± 0.2a 0.128 ± 0.04a 152.1 ± 2.1a 0.178 ± 0.02a 223.2 ± 1.4a 0.218 ± 0.01a
20 °C 138.1 ± 0.7a 0.128 ± 0.01a 153.4 ± 1.1a 0.174 ± 0.02a 223.3 ± 1.1a 0.218 ± 0.01a
30 °C 139.0 ± 1.6a 0.128 ± 0.02a 156.2 ± 1.8a 0.176 ± 0.01a 226.8 ± 3.2a 0.224 ± 0.06a
40 °C 139.5 ± 2.2a 0.129 ± 0.02a 157.6 ± 2.7a 0.175 ± 0.02a 225.5 ± 2.8a 0.214 ± 0.03a
50 °C 139.0 ± 1.4a 0.127 ± 0.01a 155.6 ± 1.7a 0.178 ± 0.03a 224.3 ± 1.4a 0.218 ± 0.01a
60 °C 138.2 ± 0.7a 0.127 ± 0.03a 157.5 ± 2.2a 0.178 ± 0.01a 227.7 ± 3.6a 0.217 ± 0.02a
70 °C 139.6 ± 1.9a 0.127 ± 0.01a 155.7 ± 1.4a 0.179 ± 0.02a 226.6 ± 2.6a 0.219 ± 0.01a
80 °C 139.4 ± 2.1a 0.126 ± 0.02a 154.8 ± 1.9a 0.178 ± 0.02a 224.3 ± 1.6a 0.219 ± 0.01a

Each value represents mean ± SD (n = 3). Tukey’s Multiple Comparison Test with a confidence interval of 95% was applied; Hd – hydrodynamic diameter; Zp – zeta potential;
PdI – polydispersity index.
a
Different letters between rows means statistically different.

against creaming after 1 year of storage. The use of small mole- Cerqueira, M., Pinheiro, A., Silva, H., Ramos, P., Azevedo, M., Flores-López, M., Rivera,
M., Bourbon, A., Ramos, Ó., Vicente, A., 2014. Design of bio-nanosystems for oral
cules’ surfactants as SDS leads to smaller droplet sizes due to the
delivery of functional compounds. Food Eng. Rev. 6 (1–2), 1–19.
faster adsorption kinetics to the interface that can reduce size Cerqueira, M.A., Bourbon, A.I., Pinheiro, A.C., Silva, H.D., Quintas, M.A.C., Antonio,
and recoalescence phenomena. The increase of surface charge A.V., 2013. Edible nano-laminate coatings for food applications. Ecosustainable
may slow down the creaming rate, due to the increase of the repul- Polymer Nanomaterials for Food Packaging. CRC Press, pp. 221–252.
Cheong, J.N., Tan, C.P., Man, Y.B.C., Misran, M., 2008. a-Tocopherol nanodispersions:
sive forces between droplets. This work showed that it is possible preparation, characterization and stability evaluation. J. Food Eng. 89 (2), 204–
to tailor charged nanoemulsions’ size through high-pressure 209.
homogenization and the appropriate choice of surfactant. Donsì, F., Sessa, M., Ferrari, G., 2011. Effect of emulsifier type and disruption
chamber geometry on the fabrication of food nanoemulsions by high pressure
homogenization. Ind. Eng. Chem. Res. 51 (22), 7606–7618.
EFSA, 2010. Call for scientific data on food additives permitted in the EU and
Acknowledgements
belonging to the functional classes of emulsifiers, stabilisers and gelling agents.
Esquena, J., Sankar, Solans, C., 2003. Highly concentrated W/O emulsions
Hélder D. Silva, and Miguel A. Cerqueira (SFRH/BD/81288/2011, prepared by the PIT method as templates for solid foams. Langmuir 19 (7),
and SFRH/BPD/72753/2010, respectively) are the recipients of a 2983–2988.
FDA, 2014a. CFR – Code of Federal Regulations Title 21.
fellowship from the Fundação para a Ciência e Tecnologia (FCT, FDA, 2014b. CFR – Code of Federal Regulations Title 21 – Food and Drugs.
Portugal). The authors would like to acknowledge to Francesco FDA, 2014c. CFR – Code of Federal Regulations Title 21 – indirect food additives:
Donsì and Giovanna Ferrari from Department of Industrial adjuvants, production aids and sanitizers.
Gordon, V., Marom, G., Magdassi, S., 2014. Formation of hydrophilic nanofibers from
Engineering, University of Salerno for helping in the viscosity and nanoemulsions through electrospinning. Int. J. Pharm. 478 (1), 172–179.
density measurements and to Rui Fernandes from IBMC, Guttoff, M., Saberi, A.H., McClements, D.J., 2015. Formation of vitamin D
University of Porto for assistance in taking the TEM pictures. The nanoemulsion-based delivery systems by spontaneous emulsification: factors
affecting particle size and stability. Food Chem. 171, 117–122.
authors thank the FCT Strategic Project High Vive, 2013. <[Link]
PEst-OE/EQB/LA0023/2013 and the project ‘‘BioInd–Biotechnolog Improveat, 2014. <[Link]
y and Bioengineering for improved Industrial and Agro-Food pro- Jadhav, A.J., Holkar, C.R., Karekar, S.E., Pinjari, D.V., Pandit, A.B., 2015. Ultrasound
assisted manufacturing of paraffin wax nanoemulsions: process optimization.
cesses’’, [Link]-07-0124- FEDER-000028, co-funded by the
Ultrason. Sonochem. 23, 201–207.
Programa Operacional Regional do Norte (ON.2 – O Novo Norte), Komaiko, J., McClements, D.J., 2015. Low-energy formation of edible nanoemulsions
QREN, FEDER. We also thank to the European Commission: by spontaneous emulsification: factors influencing particle size. J. Food Eng.
146, 122–128.
BIOCAPS (316265, FP7/REGPOT-2012-2013.1). The support of EU
Li, Y., Hu, M., Xiao, H., Du, Y., Decker, E.A., McClements, D.J., 2010. Controlling the
Cost Action FA1001 is gratefully acknowledged. Also the authors functional performance of emulsion-based delivery systems using multi-
acknowledge Stepan for providing the Neobee 1053 oil. component biopolymer coatings. Eur. J. Pharm. Biopharm. 76 (1), 38–47.
Li, Y., McClements, D.J., 2011. Inhibition of lipase-catalyzed hydrolysis of emulsified
triglyceride oils by low-molecular weight surfactants under simulated
References gastrointestinal conditions. Eur. J. Pharm. Biopharm. 79 (2), 423–431.
Life Enhancement, 2013 <[Link]
Abbas, S., Hayat, K., Karangwa, E., Bashari, M., Zhang, X., 2013. An overview of nanoresveratrol>.
ultrasound-assisted food-grade nanoemulsions. Food Eng. Rev. 5 (3), 139–157. Liu, W., Sun, D., Li, C., Liu, Q., Xu, J., 2006. Formation and stability of paraffin oil-in-
Acosta, E., 2009. Bioavailability of nanoparticles in nutrient and nutraceutical water nano-emulsions prepared by the emulsion inversion point method. J.
delivery. Curr. Opin. Colloid Interface Sci. 14, 3–15. Colloid Interface Sci. 303 (2), 557–563.
Anton, N., Gayet, P., Benoit, J.-P., Saulnier, P., 2007. Nano-emulsions and Malvern, I., 2011. Dynamic light scattering common terms defined, in: Instruments,
nanocapsules by the PIT method: an investigation on the role of the M. (Ed.), Worcestershire, UK.
temperature cycling on the emulsion phase inversion. Int. J. Pharm. 344 (1– Mason, T.G., Wilking, J.N., Meleson, K., Chang, C.B., Graves, S.M., 2006.
2), 44–52. Nanoemulsions: formation, structure, and physical properties. J. Phys.:
Anton, N., Vandamme, T., 2011. Nano-emulsions and micro-emulsions: Condens. Matter 18 (41), R635.
clarifications of the critical differences. Pharm. Res. 28 (5), 978–985. McClements, D., 2005. Food Emulsions: Principles, Practice, and Techniques, 2nd
AquaNova, 2013a. <[Link] ed. CRC Press, Boca Raton, Florida.
NovaSOL_beverage.pdf>. McClements, D.J., 2002. Lipid-Based Emulsions and Emulsifiers. CRC Press, Food
AquaNova, 2013b. <[Link] Lipids.
NovaSOL_BCS.pdf>. McClements, D.J., 2010. Emulsion design to improve the delivery of functional
Barnes, H.A., Holbrook, S.A., 1993. High concentration suspensions. In: Shamlou, P.A. lipophilic components. Ann. Rev. Food Sci. Technol. 1 (1), 241–269.
(Ed.), Preparation and Properties. Processing of Solid–Liquid Suspensions. Morais Diane, J.M., Burgess, J., 2014. Vitamin E nanoemulsions characterization and
Butterworth-Heinemann, Boston, pp. 222–245. analysis. Int. J. Pharm. 465 (1–2), 455–463.
Berton-Carabin, C.C., Ropers, M.-H., Genot, C., 2014. Lipid oxidation in oil-in-water NanoSinergy, 2013. <[Link]
emulsions: involvement of the interfacial layer. Comprehens. Rev. Food Sci. NutraLease, 2011. <[Link]
Food Safe. 13 (5), 945–977. DBID=1&LNGID=1&TMID=84&FID=767>.
Brösel, S., Schubert, H., 1999. Investigations on the role of surfactants in mechanical Ozturk, B., Argin, S., Ozilgen, M., McClements, D.J., 2014. Formation and stabilization
emulsification using a high-pressure homogenizer with an orifice valve. Chem. of nanoemulsion-based vitamin E delivery systems using natural surfactants:
Eng. Process. 38 (4–6), 533–540. Quillaja saponin and lecithin. J. Food Eng. 142, 57–63.

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]
10 H.D. Silva et al. / Journal of Food Engineering xxx (2015) xxx–xxx

Pinheiro, A.C., Lad, M., Silva, H.D., Coimbra, M.A., Boland, M., Vicente, A.A., 2013. Silva, H.D., Cerqueira, M.A., Souza, B.W.S., Ribeiro, C., Avides, M.C., Quintas, M.A.C.,
Unravelling the behaviour of curcumin nanoemulsions during in vitro Coimbra, J.S.R., Carneiro-da-Cunha, M.G., Vicente, A.A., 2011. Nanoemulsions of
digestion: effect of the surface charge. Soft Matter 9 (11), 3147–3154. b-carotene using a high-energy emulsification-evaporation technique. J. Food
Qian, C., McClements, D.J., 2011. Formation of nanoemulsions stabilized by model Eng. 102 (2), 130–135.
food-grade emulsifiers using high-pressure homogenization: factors affecting Solans, C., Solé, I., 2012. Nano-emulsions: formation by low-energy methods. Curr.
particle size. Food Hydrocolloids 25 (5), 1000–1008. Opin. Colloid Interface Sci. 17 (5), 246–254.
Rao, J., McClements, D.J., 2012. Lemon oil solubilization in mixed surfactant Solgar, 2013. <[Link]
solutions: rationalizing microemulsion & nanoemulsion formation. Food [Link]>.
Hydrocolloids 26 (1), 268–276. Song, B., Springer, J., 1996. Determination of interfacial tension from the profile of a
Rao, J., McClements, D.J., 2013. Optimization of lipid nanoparticle formation for pendant drop using computer-aided image processing: 2. Experimental. J.
beverage applications: influence of oil type, cosolvents, and cosurfactants on Colloid Interface Sci. 184 (1), 77–91.
nanoemulsion properties. J. Food Eng. 118 (2), 198–204. Sood, S., Jain, K., Gowthamarajan, K., 2014. Optimization of curcumin nanoemulsion
Sari, T.P., Mann, B., Kumar, R., Singh, R.R.B., Sharma, R., Bhardwaj, M., Athira, S., for intranasal delivery using design of experiment and its toxicity assessment.
2015. Preparation and characterization of nanoemulsion encapsulating Colloids Surf., B 113, 330–337.
curcumin. Food Hydrocolloids 43, 540–546. Stang, M., Schuchmann, H., Schubert, H., 2001. Emulsification in high-pressure
Schramm, L.L., 2006a. Dispersion and dispersed species characterization. Emulsions, homogenizers. Eng. Life Sci. 1 (4), 151–157.
Foams, and Suspensions. Wiley-VCH Verlag GmbH & Co. KGaA, pp. 13–51. Tadros, T., Izquierdo, P., Esquena, J., Solans, C., 2004. Formation and stability of
Schramm, L.L., 2006b. Interfacial energetics. Emulsions, Foams, and Suspensions. nano-emulsions. Adv. Colloid Interface Sci. 108–109, 303–318.
Wiley-VCH Verlag GmbH & Co. KGaA, pp. 53–100. Tadros, T.F., 2013. Emulsion Formation, Stability, and Rheology. Emulsion
Schramm, L.L., 2006c. Introduction. Emulsions, Foams, and Suspensions. Wiley-VCH Formation and Stability. Wiley-VCH Verlag GmbH & Co. KGaA, pp. 1–75.
Verlag GmbH & Co. KGaA, pp. 1–12. Tan, C.P., Nakajima, M., 2005. b-Carotene nanodispersions: preparation,
Shafiq, S., Shakeel, F., Talegaonkar, S., Ahmad, F.J., Khar, R.K., Ali, M., 2007. characterization and stability evaluation. Food Chem. 92 (4), 661–671.
Development and bioavailability assessment of ramipril nanoemulsion Troncoso, E., Aguilera, J.M., McClements, D.J., 2012. Influence of particle size on the
formulation. Eur. J. Pharm. Biopharm. 66 (2), 227–243. in vitro digestibility of protein-coated lipid nanoparticles. J. Colloid Interface
Shemen Industries, 2013. <[Link] Sci. 382 (1), 110–116.
potential>. Walstra, P., 1993. Principles of emulsion formation. Chem. Eng. Sci. 48 (2), 333–349.
Silva, H., Cerqueira, M., Vicente, A., 2012. Nanoemulsions for food applications: Wild Flavors Inc., 2013. <[Link]
development and characterization. Food Bioprocess Technol. 5 (3), 854–867. Colors_Emulsions_1_pdf>.

Please cite this article in press as: Silva, H.D., et al. Influence of surfactant and processing conditions in the stability of oil-in-water nanoemulsions. Journal
of Food Engineering (2015), [Link]