Available online at www.sciencedirect.

com

Food Research International 41 (2008) 172–183

www.elsevier.com/locate/foodres

Nano-particle encapsulation of fish oil by spray drying

Seid Mahdi Jafari a,*, Elham Assadpoor a, Bhesh Bhandari b, Yinghe He c
a
Department of Food Science and Technology, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran
b
School of Land and Food Sciences, University of Queensland, Brisbane, Australia
c
School of Engineering, James Cook University, Townsville, Australia

Received 26 August 2007; accepted 4 November 2007

Abstract

Nano-particle encapsulation by spray drying was undertaken by preparing sub-micron emulsions via high energy emulsifying tech-
niques, namely Microfluidization and Ultrasonication. The encapsulation efficiency of fish oil as a core material was investigated.
The attention was given to the surface oil content and surface oil coverage of encapsulated powders which are very significant parameters
in the encapsulation process. maltodextrin combined with a surface-active biopolymer (modified starch or whey protein concentrate) at a
ratio of 3:1 were used as the wall material. Results showed that Microfluidization was an efficient emulsification technique resulting in
fish oil encapsulated powder with the lowest unencapsulated oil at the surface of particles, mainly due to its capability to produce emul-
sions at the nano-range (d43 of 210–280 nm).
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Surface oil content; Nano-emulsion; Emulsification; Surface oil coverage; Microfluidization

1. Introduction Madene, Jacquot, Scher, & Desobry, 2006; Reineccius,

2001). This is intended to produce high quality encapsu-
Encapsulation is a rapidly expanding technology with a lated powders with maximum recovery. The properties of
lot of potential in different areas including pharmaceutical wall and core materials as well as the emulsion characteris-
and food industries. It is a process by which small particles tics and drying parameters are the factors that can affect
of core materials are packaged within a wall material to the efficiency of encapsulation.
form microcapsules (Gouin, 2004; Thies, 2001). One of Emulsification plays a key role in optimising the encap-
the common techniques to produce encapsulated products sulation efficiency of food flavours and oils (Liu, Furuta,
is spray drying, which involves conversion of liquid oils Yoshii, & Linko, 2000; Liu et al., 2001). It has been well
and flavours in the form of emulsions into dry powders, documented that emulsion droplet size has a pronounced
as an important application of microencapsulation in the effect on the encapsulation efficiency of different core mate-
food industry (Bhandari, 2005; Reineccius, 2004). Over rials during spray drying (Risch & Reineccius, 1988; Soot-
the last few years, the main emphasis of microencapsula- titantawat, Yoshii, Furuta, Ohkawara, & Linko, 2003;
tion of food flavours and oils has concentrated on improv- Soottitantawat et al., 2005). These reports clearly show
ing the encapsulation efficiency during spray drying, that is that reducing emulsion size can result in encapsulated pow-
preventing volatile losses and extending the shelf-life of the ders with higher retention of volatiles and lower content of
products by minimizing the amount of unencapsulated oil unencapsulated oil at the surface of powder particles. The
at the surface of powder particles (Desai & Park, 2005; presence of oil on the surface of the powder particles is
the most undesirable property of encapsulated powders.
This surface oil not only deteriorates the wettability and
*
Corresponding author. Tel./fax: +98 171 4426 432. dispersability of the powder (Millqvist-Fureby, Elofsson,
E-mail address: smjafari@gau.ac.ir (S.M. Jafari). & Bergenstahl, 2001; Vega, Kim, Chen, & Roos, 2005),

0963-9969/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodres.2007.11.002
 S.M. Jafari et al. / Food Research International 41 (2008) 172–183 173

but also is readily susceptible to oxidation and the develop- combination with maltodextrin (DE 16-20, Fieldose 17-C
ment of rancidity. Workers such as Kim, Chen, and Pearce AP, Penford Limited, NSW, Australia). Analytical grade
(2002, 2005a, 2005b) by analysing industrial dairy powders hexane and petroleum ether (BP 40–60 °C) were purchased
with X-ray photoelectron spectroscopy (XPS), have found from Sigma Chemicals Company (Sydney, Australia). Dis-
that there is a relatively high surface fat coverage on these tilled water was used for the preparation of all solutions.
powders (e.g., 53% for WPC powders). In another study, All general chemicals used in this study were of analytical
Keogh and O’Kennedy (1999) showed that milk fat encap- grade.
sulated powders with whey proteins had more than 30% fat
coverage. Some workers have also shown that the type of 2.2. Preparation of emulsions
fat has a strong influence on the level of surface fat (Kim
et al., 2005b; Millqvist-Fureby, 2003). Recently Vega and Hydrated solution of emulsion continuous phase was
Roos (2006) have made a comprehensive review on prepared by dissolving wall material powders in distilled
spray-dried dairy emulsions with an emphasize on surface water using a high speed blender (Model RW 20.n, IKA
composition. Works, Malaysia). They were produced one day before
Much of the work in this area has been done by emul- emulsification and kept overnight in a shaking water bath
sions having a droplet size of more than one micron and (Ratek Instruments, VIC, Australia) to warrant a full sat-
the application of sub-micron (nano) emulsions in encapsu- uration of the polymer molecules. For starch-based bio-
lation of oils and flavours is scant in the literature. By the polymers, the temperature of water bath was adjusted to
advent of modern emulsification systems and their poten- 60 °C while for proteins, they were kept at ambient temper-
tial application in encapsulation of food ingredients, ature to avoid changes due to temperature. In the case of
understanding the influence of emulsion size in nano-range WPC, their solutions were prepared by dispersing the
on surface oil content and coverage during spray drying is desired amount of their powder (10 wt%) into buffer solu-
essential (Jafari, Assadpoor, Bhandari, & He, 2007b). In tion (5 mM phosphate buffer, pH 7). The pH of WPC solu-
fact, there is no clear cut evidence on how sub-micron or tions was adjusted back to pH 7.0 using 1 M HCL if
nano-emulsions can improve the encapsulation efficiency required. The total concentration of dissolved solid was
of food flavours and oils. Recently, nano-emulsions have 40% (w/w) that was composed of 30 wt% maltodextrin
attracted considerable attention in various industrial fields and 10 wt% of emulsifying ingredients including one of
including cosmetics, pharmaceuticals and agrochemicals the biopolymers of Hi-Cap or WPC.
(Jafari, He, & Bhandari, 2006; Jafari, He, & Bhandari, All emulsions produced were of the oil-in-water type
2007a; Solans, Izquierdo, Nolla, Azemar, & Garcia-Celma, and being prepared in two stages: (a) pre-emulsions were
2005; Tadros, Izquierdo, Esquena, & Solans, 2004). These obtained by a rotor–stator system (Model L2R, Silverson
emulsions are kinetically stable systems with very small Machines Ltd., UK). The Silverson is a typical colloid mill
droplet sizes that can be of real benefit for encapsulation with a stator composed of a metal grating in which, 2 mm
purposes, since the stability and other features of the infeed holes are drilled. The core material (fish oil) in the ratio of
emulsion such as droplet size and distribution play a criti- 1:4 (core:wall) was progressively added to the continuous
cal role on the retention and surface oil content of the phase during pre-emulsion preparation and stirred for
product. Therefore, the objectives of this work are to deter- 10 min at the highest speed. (b) These coarse emulsions
mine the influence of sub-micron emulsions produced by were then further emulsified using a Microfluidizer (Model
different emulsification methods on encapsulation effi- M-110 L, Microfluidics, USA) at 60 MPa for one cycle, or
ciency and investigate the encapsulated powder properties an 24 KHz Ultrasound probe (Dr. Hielscher series, Model
after spray drying for different emulsion droplet sizes and UP 400S) with 22 mm diameter at the highest power for
surface-active biopolymers. In this work, we have referred 100 s. More details about emulsification conditions were
the encapsulation as ‘‘nano-particle encapsulation” since presented in our previous work (Jafari & He et al.,
the core material in nano-size range will be encapsulated 2007a; Jafari & Assadpoor et al., 2007b). Sodium azide
into the matrix of micron size powder particles. (0.02 wt%) was added to the emulsions as an antimicrobial
agent. For each emulsifying device, about 1000 mL sample
2. Materials and methods was prepared for the production of encapsulated powders
by spray drying.
2.1. Materials
2.3. Spray drying
In this study, fish oil (HiDHA 25N, Nu-Mega Ingredi-
ents, Brisbane, Australia) was used as the core material The infeed emulsions were transformed to encapsulated
(q = 850 kg/m3, g = 86 mPa s at 25 °C, RI = 1.483). The powder in a pilot-plant spray drier (Model SL 20, Saurin
wall material was an aqueous solution of a modified starch Group of Companies, Victoria, Australia). It was com-
(Hi-Cap 100, National Starch and Chemical, NSW, Aus- posed of a cylindrical chamber with a 1200 mm diameter
tralia) and/or whey protein concentrate (WPC) (ALACEN, and 2000 mm height, followed by a conical chamber of
New Zealand Milk Products, Auckland, New Zealand) in a 500 mm high with a 60° angle. The dryer had a water evap-
174 S.M. Jafari et al. / Food Research International 41 (2008) 172–183

oration rate of 3 kg/h (at 200 °C inlet/80 °C outlet) and was (Model KS 130B, IKA Works, Germany) and extraction
equipped with a pressure air atomizing nozzle and auto- was extended for 15 min. After the required extraction
mated feeding of the liquid into drying chamber through time, the powder and the solvent were separated by filtra-
a peristaltic pump. Feed and hot air were entering the spray tion through filter paper (No. 41, Whatman, Maidstone,
drier in a co-current manner. The operational conditions of UK). The powder residue was washed with 3  5 mL sol-
spray drying (obtained previously) were: air inlet tempera- vent and the filtrate solution containing the extracted oil
ture of 180 °C, air outlet temperature of 65 °C, and nozzle was transferred to a round-bottomed flask, which was sub-
air pressure of 310 kPa. The dried powder was collected sequently placed in a rotary evaporator (Model R-114,
and stored in opaque, air tight containers at 4 °C awaiting Buchi Co., Flawil, Switzerland) to evaporate the organic
for further analysis. solvent under vacuum condition. The flask after evapora-
tion was dried in an oven (IM550, Clayson Inc., NSW,
2.4. Emulsion droplet size and powder particle size analysis Australia) at 103 °C until constant weight (about 1 h)
and the extracted oil value was then calculated based on
The size distribution of oil droplets in infeed emulsions the difference between the weight of the initial clean flask
was determined by the laser light scattering method using and that containing extracted oil residue.
Mastersizer 2000 (Malvern Instruments, Worcestershire,
UK), and reported in d32 or d43 as described in our previ- 2.7. Surface oil coverage analysis
ous study (Jafari & He et al., 2007a; Jafari & Assadpoor
et al., 2007b). In order to investigate emulsion size change Because of the possible extraction of core material from
during the encapsulation process, droplet size of the recon- internal parts of the encapsulated particles during solvent
stituted emulsion prepared from the encapsulated powders extraction, we analysed the surface composition of the fish
was determined by Mastersizer 2000. Redispersion proper- oil encapsulated powders by X-ray photoelectron spectros-
ties were studied by producing 10% (w/w) aqueous solu- copy (XPS). This is a well-established technique for ele-
tions of the powders by mixing the reconstituted mental analysis of the surface composition of materials.
emulsions at room temperature with a magnetic stirrer The underlying principle of the technique and its applica-
for 30 min. The analysis of powder particle size was per- tion to food powders, particularly dairy powders, have
formed using the laser light scattering method by an ana- been described elsewhere (Faldt, Bergenstahl, & Carlsson,
lyser with a batch cell unit (Mastersizer E, Malvern 1993; Gaiani et al., 2006; Keogh & O’Kennedy, 1999;
Instruments, Worcestershire, UK). Encapsulated powders Kim et al., 2002, 2005a; Keogh et al., 2001). The XPS anal-
were dispersed in propan-2-ol for the particle size analysis. ysis was carried out with a Kratos Axis Ultra (Kratos Ana-
In all cases, volume-mean diameter or d43 was reported as lytical, Manchester, UK) spectrometer incorporating a
the emulsion size or powder particle size. 165 mm hemispherical electron energy analyser. The inci-
dent radiation was monochromatic Al X-rays (1486.6 eV)
2.5. Encapsulation efficiency analysis at 150 W (15 kV, 15 ma). Survey (wide) scans were taken
at an analyser pass energy of 160 eV and multiplex (nar-
One of the important parameters for encapsulated pow- row) high resolution scans at 20 eV. Survey scans were car-
ders is the encapsulation efficiency during the process. By ried out over 1200–0 eV binding energy range with 1.0 eV
definition, it is the amount of core material encapsulated steps and a dwell time of 100 ms. Narrow high resolution
inside the powder particles. Accordingly, surface oil has scans were run with 0.05 eV steps and 250 ms dwell time.
been named as free or extractable oil, as it is measured Base pressure in the analysis chamber was 1.0  107 Pa
mostly through solvent extraction. In this study, no total and during sample analysis 1.3  106 Pa.
oil analysis was performed assuming that all the initial oil Fish oil encapsulated powders contain many compo-
was retained in the powder because of being non-volatile nents, but in order to perform the XPS measurement, it
and depositions of the unbounded oil on the dryer wall is considered that the powder is composed of only three
and degradation of oil during spray drying was ignored. main components (starch, protein, oil). In these particular
Instead, surface oil content or surface oil coverage were experiments, the relative atomic concentration (%) of car-
selected criteria for encapsulation efficiency. bon, oxygen and nitrogen in the surface layer (10 nm)
of the powder sample was analysed. Elemental composi-
2.6. Surface oil content measurements tion in the sample is assumed to be a linear combination
of the elemental composition of the pure components mak-
The surface oil content of fish oil encapsulated powders ing up samples. The relative elemental composition of the
was determined using extraction with petroleum ether (bp pure components measured by XPS and used in the calcu-
40–60 °C), according to the method of Garcia, Gutierrez, lations is shown in Table 1. The values are an average of
Nolasco, Carreon, and Arjona (2006). Spray dried powder two independent analyses. The percent composition (rela-
(10 g) was accurately weighed and added to 100 mL of sol- tive coverage) of the different components such as fish oil
vent in a volumetric flask and the suspension was shaken in the powder surface layer was calculated by using the
frequently by hand. Then, it was placed on a shaker relation in a matrix formula as
 S.M. Jafari et al. / Food Research International 41 (2008) 172–183 175

Table 1 OE and NE), starch including Hi-Cap and maltodextrin (CS,

Relative elemental composition of pure encapsulation ingredients mea- OS and NS), protein, i.e., WPC (CP, OP and NP), and finally
sured by XPS
fish oil (CO, OO, NO); which are values obtained from the
Component Relative atomic concentration (%) areas of the C 1s, O 1s, and N 1s at the XPS peaks. By solv-
C O N ing the matrix (MATLAB Version 4.0, 2005), values of a, b,
Maltodextrin 58.63 ± 0.1 41.37 ± 0.1 0 and c can be obtained which are corresponding to the sur-
Hi-Cap 63.34 ± 0.2 36.62 ± 0.2 0 face starch, protein and oil contents. Since there is a small
Whey protein concentrate 77.05 ± 1.2 14.67 ± 0.8 8.29 ± 0.5 difference between the molecular weights of carbon (12),
Fish oil 92.04 ± 0.2 7.97 ± 0.2 0
oxygen (16) and nitrogen (14), the relative coverage calcu-
lated here is expected to be close to the mass based one.
CE ¼ aCS þ bCP þ cCO In Fig. 1, the survey scan of fish oil encapsulated powder
OE ¼ aOS þ bOP þ cOO containing WPC is presented as an example with the de-
NE ¼ aNS þ bNP þ cNO tailed peaks of each element. It should be mentioned that
for fish oil, the pre-operating temperature of the machine
where C, O, and N are the relative atomic concentrations of was taken down to transform the liquid oil into a frozen so-
carbon, oxygen, and nitrogen in encapsulated powder (CE, lid, and then surface analysis was performed.

Fig. 1. Example of XPS spectra obtained for fish oil (20 wt%) encapsulated powder containing WPC and maltodextrin in the proportion of 1:3. (1) Survey
scan. (2) Narrow spectra of O 1s, N 1s, and C 1s levels, respectively. (3) Mass concentrations of different decomposed peaks.
176 S.M. Jafari et al. / Food Research International 41 (2008) 172–183

2.8. Analysis of moisture content and water activity of was considered very significant at P < 0.01. Some of the
powders graphs were drawn by Excel (Microsoft Office 2003) and
some by MINITAB 14.
Moisture content of the encapsulated powders was
determined gravimetrically by oven drying (Vord-460-D, 3. Results and discussion
Thermoline Scientific Equipment Pty. Ltd., NSW, Austra-
lia) at 70 °C to constant weight. The sample size was 5.0 g 3.1. Encapsulation efficiency for different emulsification
powder widened inside glass plates and the vacuum drying systems
time was 72 h. The results were reported on the wet basis as
(weight loss/sample weight)  100. The water activity of In order to investigate the influence of emulsification sys-
spray dried powders was measured by an Aqualab system tem on the properties of fish oil (20% w/w) encapsulated
(Model 3 TE, Decagon Devices Inc., Washington, USA). powder such as the surface oil content and coverage, we pre-
pared the infeed emulsions for spray drying by three differ-
2.9. Scanning electron microscopy of encapsulated powders ent emulsifying devices including Silverson, Microfluidizer,
and Ultrasound. The infeed emulsions were processed in
A JSM 6400F model scanning electron microscope similar spray drying conditions. The wall material was a
(JEOL Co. Ltd., Tokyo, Japan) was used to investigate combination of maltodextrin and either Hi-Cap or WPC
the microstructural properties of the spray dried encapsu- as the surface-active biopolymer in the ratio of 3:1. Statisti-
lated powders. The samples were placed on the SEM stubs cal analysis results are presented in Table 2 along with final
using a two-sided adhesive tape (Nisshin EM Co. Ltd., results in Table 3.
Tokyo, Japan). The specimens were subsequently coated The overall result was that emulsification method had a
with Pt using a magnetron sputter coater (Model EIKO very significant influence (P < 0.01) on the encapsulation
IB-5, Eiko Inc., Tokyo, Japan). The coated samples were efficiency and other measured data such as emulsion size
then analysed using the SEM operating at an accelerating before spray drying and after reconstituting the encapsu-
voltage of 15 kV. The micrographs representing the micro- lated powder in water. For example, by using Hi-Cap as
structure of the encapsulated powders were taken by the the emulsifier and producing emulsion with Microfluidizer,
instrument’s software installed on a PC connected to the surface oil content of fish oil encapsulated powders was
system. reducing very significantly (P < 0.01) to about 170 mg/
100 g powder compared with Silverson samples (2040 mg/
2.10. Experimental design and statistical analysis 100 g powder) or Ultrasound samples (540 mg/100 g pow-
der). This can be explained by parameters such as the size
The parameters considered affecting the encapsulation and distribution of infeed emulsion, stability of emulsions
efficiency were the emulsification method, the type of sur- during spray drying, powder particle size, and the influence
face-active biopolymer, and emulsion droplet size. The of the emulsification itself on the emulsion properties. As it
effect of each operational and compositional parameter can be seen from Table 3, fish oil emulsions produced with
on the encapsulation efficiency (surface oil content and Silverson had a significantly bigger size (P < 0.05) than
coverage) was studied using only one system (emulsifying with Microfluidizer and Ultrasound. For both Hi-Cap
method or wall material). All the experiments were per- and WPC, Microfluidized emulsions had the lowest droplet
formed based on a fully factorial design and the results rep- size (210 and 280 nm, respectively) while, the same emul-
resent the means of two replicates. General Linear Model sions produced with Silverson had a droplet size of about
of MINITAB (Version 14, 2004) was used to conduct an 4.6 and 5.9 lm for Hi-Cap and WPC samples, respectively.
analysis of variance (ANOVA) to determine differences The reason could be high viscosity of fish oil that is not in
between treatments means. Treatments means were consid- the optimum range of rotor–stator emulsification (Walstra,
ered significantly different at P 6 0.05 and the difference 2003).

Table 2
The P-values obtained with ANOVA realised on the variables versus different factors affecting encapsulation efficiency of fish oil
Response/factor Emulsion size Powder size Reconstituted Surface oil (based Surface oil (based Surface oil
(d43) (d43) emulsion size (d43) on weight) on surface area) coverage (%)
A: Blocks 0.148 0.096 0.355 0.110 0.055 0.142
(replications)
B: Emulsifier (WPC 0.001 VS 0.118 0.001 VS 0.056 0.188 0.001 VS
vs. Hi-Cap)
C: Emulsification 0.001 VS 0.020 S 0.001 VS 0.001 VS 0.001 VS 0.001 VS
method
B:C (interaction) 0.001 VS 0.519 0.001 VS 0.297 0.119 0.001 VS
Significance levels: VS = very significant (P < 0.01); S = significant (P < 0.05).
 S.M. Jafari et al. / Food Research International 41 (2008) 172–183 177

Table 3
Influence of the emulsifying method and emulsifier type on surface oil content and other properties of fish oil encapsulated powders
Surface active Emulsification Emulsion size Powder size Reconstituted Powder moisture Surface oil Surface oil Surface oil
biopolymer method (d43, lm) (d43, lm) emulsion size (wt%) (mg/100 g powder) (mg/m2 powder) coverage (%)
(d43, lm)
Hi-Cap Silverson 4.6a 41.2a 48.5a 1.4a 2040a 355.4a 58.1a
Microfluidizer 0.21b 25.7b 0.26b 1.7b 170b 22.2b 18.9b
Ultrasound 2.2c 28.8b 4.1c 1.7b 540c 74.5c 25.6c
WPC Silverson 5.9a 36.0a 6.8a 1.6a 1270a 174.8a 35.2a
Microfluidizer 0.28b 32.1a 0.88b 1.3b 690b 98.3b 18.3b
Ultrasound 3.5c 34.0a 2.5c 1.5a,b 770b 103.6b 24.1c
Means within the same column (for individual biopolymers) followed by different letters (a, b, c) are significantly (P < 0.05) different.

Small oil droplets will be enclosed and embedded more particle size except Silverson samples (Table 3). For
efficiently within the wall matrix of the microcapsules and instance, surface oil content of Microfluidized samples con-
also, the resulted emulsion will be more stable during the taining Hi-Cap and WPC were 170 and 690 mg/100 g pow-
spray drying encapsulation process which is one of the crit- der, respectively. Representing the surface oil data based
ical parameters to have the optimum efficiency. Therefore, on specific surface area of powder particles could not
it is expected that Microfluidized emulsions will lead to change the trend as there was a small difference between
encapsulated powders with the minimum amount of unen- particle size of Hi-Cap and WPC powders. This could be
capsulated oil (surface oil). Our data confirmed this trend, explained by lower stability of WPC stabilized emulsions
particularly with samples containing Hi-Cap (Fig. 2). Dan- during spray drying process compared with Hi-Cap emul-
viriyakul, McClements, Decker, Nawar, and Chinachoti sions as droplet size of redispersed WPC emulsions showed
(2002) reported similar results by spray drying milk fat a shift towards bigger size (Fig. 3). In fact, there was more
emulsions with sodium caseinate and corn syrup solids coalescence in WPC samples than in Hi-Cap emulsions
with lecithin. They found that the level of extractable fat inferring that some of WPC molecules will lose their emul-
(surface fat) increased significantly (P < 0.05) with increas- sifying and stabilizing abilities during the spray drying
ing emulsion size from 0.5 lm (2% surface fat) to 1.2 lm encapsulation process.
(13% surface fat), due to instability (creaming) of emul- Surprisingly, the surface oil content of WPC samples
sions with bigger droplets. from both Ultrasound and Microfluidized emulsions were
comparable with no significant difference, although Ultra-
3.2. Encapsulation efficiency for different surface-active sound emulsions had a much higher droplet size (d43 =
biopolymers 3.5 lm). The powder particle size for both of these samples
was similar and observation of their microstructure with
When comparing fish oil encapsulated powders consist- SEM also did not reveal any difference in their structures
ing of Hi-Cap with those made with WPC, it was found (results not shown). The only reason for this unexpected
that Hi-Cap samples have generally lower amounts of sur- behaviour could be stability of the produced emulsions
face oil with marginally smaller emulsion size and powder during encapsulation spray drying. It was found that drop-
let size (d43) of Microfluidized fish oil emulsions made with
WPC increased almost 4 times from 0.28 to 0.88 lm during
the process, while d43 of their Ultrasound counterparts
decreased 30% from 3.5 to 2.5 lm. possibly because of
atomization and reconstitution conditions. In other words,
Surface oil (mg/100 g powder)

2250
Ultrasound emulsions have undergone fewer changes than
1800 Microfluidized ones suggesting that emulsification method
can be important as different results obtained. This issue
1350 needs to be further investigated.
Another important result was different stability of fish
900
oil emulsions containing Hi-Cap during spray drying. For
450
WPC instance, initial Silverson emulsions made with Hi-Cap
Hi-Cap
had a d43 of about 4.6 lm and after spray drying, reconsti-
0 tuted emulsions made from encapsulated powders had a
Silverson Ultrasound Microfluidizer bimodal distribution with a d32 = 0.94 lm, and d43 =
Emulsifying device 48.5 lm. In the same conditions, d43 of Microfluidized sam-
Fig. 2. Surface oil content of fish oil (20 wt%) encapsulated powders ples only increased slightly from 0.21 lm (initial emulsion)
containing two different surface-active biopolymers (WPC and Hi-Cap) to 0.26 lm (reconstituted emulsions), suggesting very sig-
and from emulsions of three different emulsification techniques. nificant stability of Microfluidized emulsions (P < 0.05)
178 S.M. Jafari et al. / Food Research International 41 (2008) 172–183

A 10.0 B 10

7.5
Volume (%)

Volume (%)
5.0 5

2.5

0.0 0

100 1000 10000 100 1000 10000

Emulsion droplet size (nm) Emulsion droplet size (nm)

Fig. 3. Droplet size distribution of original (solid line) and redispersed (dashed-line) emulsions containing Hi-Cap (A) or whey protein concentrate (B) as
the emulsifying agent. Core material was fish oil (20 wt%) and emulsification was done by Microfluidizer.

compared with very poor stability of Silverson emulsions. vent (petroleum ether in our case) can reach the interior
The main reason for this could be much lower droplet size through cracks and pores at the surface of particles. So,
of Microfluidized emulsions consisted of Hi-Cap, and surface fat is the only fraction of fat that is present at the
thereby, minimum influence of atomization and spray dry- surface of encapsulated particles, mainly responsible for
ing process on the emulsions size, and as a result, substan- oxidation, and solvent extraction data does not necessarily
tially lower surface oil contents for these samples, as shown correspond to this parameter.
in Fig. 2. It should be noted that we did not analyse the We analysed the surface composition of fish oil encapsu-
total oil content (retention) of powders consisted of fish lated powders by X-ray photoelectron spectroscopy (XPS)
oil, since it is not a volatile compound and it is assumed in order to compare the results with surface oil data
that there would be no loss during the spray drying encap- obtained by extraction measurements. The comparison of
sulation process. these data could clearly show how much of the surface of
Considering particle size of fish oil encapsulated pow- powder particles is really covered by unencapsulated oil
ders, there was no significant difference (P > 0.05) between and how much oil has come to the surface through cracks
samples from various emulsification systems for WPC and and holes at the surface of powder particles. Surface fat
Hi-Cap, except Silverson samples containing Hi-Cap which coverage of fish oil encapsulated powders obtained from
had bigger particle size (41.2 lm) than other samples. So, it different emulsification techniques and containing different
can be concluded that powder particle size does not play an surface active ingredients is shown in Fig. 4.
important role on the overall encapsulation efficiency of It was found that emulsification technique has a signifi-
fish oil emulsions. Finney, Buffo, and Reineccius (2002) cant influence (P < 0.05) on surface oil coverage of fish oil
could also find no effect of particle size on retention of encapsulated powders, irrespective of the surface-active
some volatiles, as they attributed this result to the high con- biopolymer, and similar to the surface oil content data
centration of infeed solids, i.e., particle size is not impor-
tant if high infeed solids are used.

3.3. Surface oil coverage and emulsification systems

60
Surface oil coverage (%)

Surface oil data should be treated with cautions due to

50
possible removal of some encapsulated oil from internal
part of the microcapsules through solvent extraction anal- 40
ysis. Different terms for surface oil are used in the literature
30
such as free fat, extractable fat, unencapsulated or surface
fat (oil) with an aim to measure and represent the fraction 20
WPC
of oil present at the surface of encapsulated powder parti-
10
cles. Free fat, defined as the amount of fat that can be Hi-Cap
extracted from encapsulated powders by an organic sol- 0
Silverson Ultrasound Microfluidizer
vent, is supposed to represent the fat present on the surface
of the powder, although it has been shown that the free fat Emulsifying device
originates not only from the surface of particles, but also Fig. 4. Surface fish oil coverage of encapsulated powders analysed by
from the interior of the powder particles. The organic sol- XPS.
 S.M. Jafari et al. / Food Research International 41 (2008) 172–183 179

(Fig. 2). Powders produced from Microfluidized emulsions ited emulsifying ability of modified starches which are
had the minimum surface oil coverage (approximately mainly related to incorporated side chains of succinic acid,
18%), while emulsions made with rotor–stator system (Silv- while in WPC, there are many hydrophobic and hydro-
erson) lead to a powder with high surface oil coverage (more philic sites having emulsifying capabilities.
than 35%). Ultrasound samples had the intermediate Workers such as Kim et al. (2002, 2005a, 2005b) by
amounts of surface oil coverage between Microfluidized analysing industrial dairy powders with XPS, have found
and Silverson samples. This trend could be well explained that there is a relatively high surface fat coverage on these
by the droplet size of initial emulsions fed into the spray powders (e.g., 53% for WPC powders). In another study,
drier, similar to the results of Keogh et al. (2001). These Keogh and O’Kennedy (1999) showed that milk fat encap-
workers found that increasing the homogenization pressure sulated powders with whey proteins had more than 30% fat
or number of passes during emulsification of fish oil samples coverage. Our results for fish oil surface coverage were rel-
decreased the surface oil coverage significantly from 47.4% atively lower than these figures (except Silverson samples),
to 37.5% (by using sodium caseinate as the wall material). possibly due to more efficient encapsulation of emulsions
Our results (Fig. 5) revealed that there was a direct rela- with smaller droplets. Also, it has been shown that the
tionship between original emulsion droplet size and surface liquid oils (e.g., fish oil) and fully crystalline fats with high
oil coverage of the encapsulated powders (r2 equal to 0.87 melting points were well encapsulated with low surface oil
and 0.97 for Hi-Cap and WPC samples, respectively): smal- coverage, while fats with intermediate melting points were
ler the emulsion size, lower the surface oil coverage. The poorly encapsulated with the highest surface fat coverage
same trend was existed for surface oil content obtained (Faldt et al., 1993; Gaiani et al., 2006; Kim et al., 2005b).
by solvent extraction with r2 equal to 0.89 and 0.85 for Unfortunately, published papers regarding surface oil cov-
powders containing Hi-Cap and WPC, respectively. In erage of encapsulated powders are scant in the literature
fact, XPS results were corresponding well with solvent and we could not compare our results to presented data
extraction data. Another important result was a higher rate other than milk fat.
of surface oil increase for Hi-Cap samples with increasing
emulsion size compared with their WPC counterparts, sug- 3.4. Surface oil coverage and emulsifying biopolymer
gesting that emulsifying and encapsulating capabilities of
Hi-Cap are limited and lower than WPC, particularly at When comparing the surface oil coverage data for two
higher emulsion sizes. This could be attributed to the lim- different surface-active biopolymers, it was found that

60 2500
Surface oil content (mg/100 g
Surface oil coverage (%)

2000
45

1500
powder)

30
1000

15
500

0 0
WPC emulsions: 0.28 3.5 5.9
Hi-Cap emulsions: 0.21 2.2 4.6
D43 (µm)
Surface oil content (mg/100 g

60 2100
Surface oil coverage (%)

y = 19.6x- 5 y = 935x- 953.33

45 1600
R2= 0.8738 R2= 0.8915
powder)

30 1100
y = 8.45x+ 8.9667 y = 290x+ 330
15 R2= 0.9683 600 R2= 0.8512

0 100
Microfluidizer Ultrasound Silverson Microfluidizer Ultrasound Silverson
Emulsification method Emulsification method

Fig. 5. Relationship between emulsion size (d43) or emulsification method and surface oil content and/or coverage of fish oil encapsulated powders
containing Hi-Cap (j, h), or WPC (N, D) as the emulsifier. Closed symbols represent surface oil coverage analysed by XPS and open symbols represent
surface oil content obtained by solvent extraction. Powder composition was fish oil (20 wt%) and maltodextrin with each emulsifiers in the ratio of 3:1.
180 S.M. Jafari et al. / Food Research International 41 (2008) 172–183

encapsulated powders containing WPC have generally that increasing the proportion of ‘‘whey protein iso-
lower surface oil coverage than their Hi-Cap counterparts, late:maltodextrin” from 1:19 to 3:1 lead to powder particles
particularly for Silverson samples. This was an important with minimum indentations. With powder particles con-
result, since our data for surface oil content obtained with taining Hi-Cap, the SEM photos (Fig. 6C and D) showed
extraction (Fig. 2) showed that WPC samples had usually that there was a fast crust formation (or solidification) that
more surface oil contents, which was very obvious for could be associated with low levels of surface oil content
Microfluidized samples. While XPS data revealed that obtained with solvent extraction as there is less opportunity
there is no significant difference (P > 0.05) between WPC for the core material (fish oil) droplets to come onto the
and Hi-Cap powders (with surface oil coverage of 18.3 surface of particles. While in WPC samples, crust forma-
and 18.9%, respectively) obtained from Microfluidized tion is slower inferring that more oil droplets could come
emulsions. This could be related to the extraction of inter- onto the surface and our results for surface oil data with
nal oil (unencapsulated or not efficiently encapsulated) extraction supports this hypothesis.
from inside the powder particles containing WPC or the Results with XPS analysis however showed no difference
existence of thick layers of oil at the surface of these parti- between WPC and Hi-Cap samples considering surface oil
cles. Observations with SEM confirmed the latter as no evi- coverage that could be attributed to the thicker layer of
dence of cracks or holes was found in WPC samples. surface oil on particles containing WPC compared with
It was found that powder particles containing WPC had thinner layer on Hi-Cap samples. In other words, the sur-
less surface dents and shrinkage than their Hi-Cap counter- face area occupied by fish oil on WPC samples is smaller
parts (Fig. 6). In other words, incorporating WPC into (but thicker) than Hi-Cap samples. These are in agreement
emulsion composition had a profound influence on the with the results of Faldt and Bergenstahl (1996) and Faldt
structure and surface morphology of encapsulated pow- et al. (1993) who concluded that samples with low extracted
ders, resulting in particles with smooth surfaces and less fat but high surface coverage of fat (such as Microfluidized
indentations. This suggested a slower rate of wall matrix Hi-Cap samples in our case) have a well encapsulated fat in
solidification with WPC samples, and possibly a higher the interior and the powders probably have a structure with
elasticity of these wall systems. These are in agreement with separate fat pools in the interior and with the surface cov-
the results of Sheu and Rosenberg (1995, 1998) who found ered with a layer of fat.

Fig. 6. SEM of fish oil encapsulated powders produced from Microfluidized emulsions: WPC samples (A and B) compared with Hi-Cap samples (C
and D).
 S.M. Jafari et al. / Food Research International 41 (2008) 172–183 181

Fig. 7. SEM micrographs of fish oil encapsulated powders containing Hi-Cap and produced from Silverson emulsions: unwashed samples (A and B)
compared with washed samples (C and D).

The importance of the mechanical properties of the wall stages of drying and the particles’ surface had not been
materials in affecting the structures of microcapsules was smoothed out at the late stages of drying. This could be
highlighted by the existence of cracks in capsules containing due to no occurrence of ballooning and no existence of
Hi-Cap and obtained from Silverson emulsions (Fig. 7). In cracks. While in Silverson samples, because of the presence
these situations, although different extents of dent smooth- of cracks, they had more spherical particles. Therefore, it
ing were observed, the wall system was too fragile to with- can be concluded that the formation of surface indenta-
stand the mechanical forces associated with expansion tions in spray dried powders not only depends on the com-
and ballooning. Relatively small proportion of microcap- position of wall materials and drying parameters, but also
sules with ‘‘caps within dents” features, particularly for on the emulsification method used to prepare infeed emul-
big particles, and a large proportion of dent-free capsules sions and more precisely, emulsion size.
(mostly smaller particles) was observed in Silverson samples
containing Hi-Cap (Fig. 7A and B). The external structure
of the powder particles containing Hi-Cap and made with 4. Conclusion
Silverson emulsions were very porous and had many deep
dents and cracks, as shown in Fig. 7. Most of these cracks The findings of our study clearly showed that emulsifica-
were located on ‘‘caps within dents” sites (confirming the tion method can influence and determine the final proper-
poor mechanical withstanding of the wall system during ties of the encapsulated powder in a number of different
expansion) and were not observable on the original powders ways including emulsion size, emulsion stability, powder
(Fig. 7A and B), as the particles had a smooth surface. But particle size, size distributions, and other parameters such
washing the powders with solvent (petroleum ether) as the surface oil of the spray dried powders. Our results
revealed the topography of the particles (porous with dents) revealed that Microfluidization was the best emulsification
and the existence of cracks and fractures that certainly method to achieve minimum amounts of unencapsulated
would be the main reason of having much higher unencap- oil at the surface of particles and maximum encapsulation
sulated oils at the surface of these particles. efficiency. Ultrasound was in the next order and obviously,
It is interesting to note that the same emulsion composi- Silverson rotor–stator emulsification was the poorest
tion emulsified with Microfluidizer and spray dried in iden- method for the preparation of infeed emulsions.
tical conditions had a totally different structure (Fig. 6C The variation of results for different emulsifying devices
and D). They had extreme shrinkage possibly in earlier could be well explained by the droplet size of produced
182 S.M. Jafari et al. / Food Research International 41 (2008) 172–183

emulsions. Smaller the emulsion size, better the encapsula- Gouin, S. (2004). Microencapsulation: Industrial appraisal of existing
tion efficiency. Although the emulsification system can also technologies and trends. Trends in Food Science & Technology, 15(7-8),
330–347.
affect other parameters such as powder particle size, but Jafari, S. M., He, Y., & Bhandari, B. (2006). Nano-emulsion production
more works need to be done on the influence of each emul- by sonication and microfluidization – A comparison. International
sification technique on the molecular structure of surface- Journal of Food Properties, 9(3), 475–485.
active biopolymers and their emulsifying capabilities, Jafari, S. M., He, Y., & Bhandari, B. (2007a). Optimization of nano-
because it was found that even two emulsions with same emulsions production by microfluidization. European Food Research
and Technology, 225, 733–741.
size from two different emulsifying devices will result in dif- Jafari, S.M., Assadpoor, E., Bhandari, B., & He, Y. (2007). Re-
ferent encapsulation efficiencies. Considering two different coalescence of emulsion droplets during high-energy emulsification.
surface-active biopolymers, we found that Hi-Cap was bet- Food Hydrocolloids, doi:10.1016/j.foodhyd.2007.09.006.
ter than WPC since it lead to encapsulated powders with Keogh, M. K., & O’Kennedy, B. T. (1999). Milk fat microencapsula-
less surface oil. Interaction of wall materials with encapsu- tion using whey proteins. International Dairy Journal, 9(9), 657–663.
Keogh, M. K., O’Kennedy, B. T., Kelly, J., Auty, M. A., Kelly, P. M.,
lated core materials is another factor which needs to be Fureby, A., et al. (2001). Stability to oxidation of spray-dried fish oil
investigated. Also, it was explained that surface oil data powder microencapsulated using milk ingredients. Journal of Food
should be treated with caution as there is a chance for Science, 66(2), 217–224.
internal oil within the microcapsules to come to the surface Kim, E. H. J., Chen, X. D., & Pearce, D. (2002). Surface characterization
by solvent extraction. XPS analysis of fish oil encapsulated of four industrial spray-dried dairy powders in relation to chemical
composition, structure and wetting property. Colloids and Surfaces B –
powders showed that different biopolymers (Hi-Cap and Biointerfaces, 26(3), 197–212.
WPC) resulted in almost similar encapsulated powders in Kim, E. H. J., Chen, X. D., & Pearce, D. (2005a). Effect of surface
terms of surface oil coverage, and here again; emulsion size composition on the flowability of industrial spray-dried dairy powders.
was significantly affecting the final results. In fact, there Colloids and Surfaces B – Biointerfaces, 46(3), 182–187.
was a strong correlation between emulsion droplet size Kim, E. H. J., Chen, X. D., & Pearce, D. (2005b). Melting characteristics
of fat present on the surface of industrial spray-dried dairy powders.
and surface oil coverage of encapsulated powders. To this Colloids and Surfaces B – Biointerfaces, 42(1), 1–8.
end, we provided SEM micrographs of encapsulated pow- Liu, X. D., Atarashi, T., Furuta, T., Yoshii, H., Aishima, S., Ohkawara,
ders containing various surface-active ingredients obtained M., et al. (2001). Microencapsulation of emulsified hydrophobic
from diverse emulsification methods and discussed that it is flavors by spray drying. Drying Technology, 19(7), 1361–1374.
a powerful and helpful system to investigate different Liu, X. D., Furuta, T., Yoshii, H., & Linko, P. (2000). Retention of
emulsified flavor in a single droplet during drying. Food Science and
behaviours and results. Technology Research, 6(4), 335–339.
Madene, A., Jacquot, M., Scher, J., & Desobry, S. (2006). Flavour
encapsulation and controlled release – A review. International Journal
References of Food Science and Technology, 41(1), 1–21.
Millqvist-Fureby, A., Elofsson, U., & Bergenstahl, B. (2001). Surface
Bhandari, B. (2005). Spray drying – An encapsulation technique for food composition of spray-dried milk protein-stabilized emulsions in
flavors. In A. S. Mujumdar (Ed.), Drying of products of biological relation to pre-heat treatment of proteins. Colloids and Surfaces B:
origin enfield. USA: Science Publishers. Biointerfaces, 21, 47–58.
Danviriyakul, S., McClements, D. J., Decker, E., Nawar, W. W., & Millqvist-Fureby, A. (2003). Characterisation of spray-dried emulsions
Chinachoti, P. (2002). Physical stability of spray-dried milk fat with mixed fat phases. Colloids and Surfaces B: Biointerfaces, 31,
emulsion as affected by emulsifiers and processing conditions. Journal 65–79.
of Food Science, 67(6), 2183–2189. Reineccius, G. A. (2001). Multiple-core encapsulation: The spray drying
Desai, K. G. H., & Park, H. J. (2005). Recent developments in of food ingredients. In P. Vilstrup (Ed.), Microencapsulation of food
microencapsulation of food ingredients. Drying Technology, 23(7), ingredients. Leatherhead Food RA Publishing.
1361–1394. Reineccius, G. A. (2004). The spray drying of food flavours. Drying
Faldt, P., & Bergenstahl, B. (1996). Spray-dried whey protein/lactose/ Technology, 22(6), 1289–1324.
soybean oil emulsions.1. Surface composition and particle structure. Risch, S. J., & Reineccius, G. A. (1988). Spray-dried orange oil – Effect of
Food Hydrocolloids, 10(4), 421–429. emulsion size on flavor retention and shelf stability. ACS Symposium
Faldt, P., Bergenstahl, B., & Carlsson, G. (1993). The surface coverage of Series, 370, 67–77.
fat on food powders analyzed by Esca (electron-spectroscopy for Sheu, T. Y., & Rosenberg, M. (1995). Microencapsulation by spray-drying
chemical-analysis). Food Structure, 12(2), 225–234. ethyl caprylate in whey-protein and carbohydrate wall systems. Journal
Finney, J., Buffo, R., & Reineccius, G. A. (2002). Effects of type of of Food Science, 60(1), 98–103.
atomization and processing temperatures on the physical properties Sheu, T. Y., & Rosenberg, M. (1998). Microstructure of microcapsules
and stability of spray-dried flavors. Journal of Food Science, 67(3), consisting of whey proteins and carbohydrates. Journal of Food
1108–1114. Science, 63(3), 491–494.
Gaiani, C., Ehrhardt, J. J., Scher, J., Hardy, J., Desobry, S., & Solans, C., Izquierdo, P., Nolla, J., Azemar, N., & Garcia-Celma, M. J.
Banon, S. (2006). Surface composition of dairy powders observed (2005). Nano-emulsions. Current Opinion in Colloid & Interface
by X-ray photoelectron spectroscopy and effects on their rehydra- Science, 10(3–4), 102–110.
tion properties. Colloids and Surfaces B – Biointerfaces, 49(1), Soottitantawat, A., Bigeard, F., Yoshii, H., Furuta, T., Ohkawara, M., &
71–78. Linko, P. (2005). Influence of emulsion and powder size on the stability
Garcia, E., Gutierrez, S., Nolasco, H., Carreon, L., & Arjona, O. (2006). of encapsulated D-limonene by spray drying. Innovative Food Science &
Lipid composition of shark liver oil: Effects of emulsifying and Emerging Technologies, 6(1), 107–114.
microencapsulation processes. European Food Research and Technol- Soottitantawat, A., Yoshii, H., Furuta, T., Ohkawara, M., & Linko, P.
ogy, 222(5–6), 697–701. (2003). Microencapsulation by spray drying: Influence of emulsion size
 S.M. Jafari et al. / Food Research International 41 (2008) 172–183 183

on the retention of volatile compounds. Journal of Food Science, 68(7), Vega, C., Kim, E. H. J., Chen, X. D., & Roos, Y. H. (2005). Solid-state
2256–2262. characterization of spray-dried ice cream mixes. Colloids and Surfaces
Tadros, T., Izquierdo, R., Esquena, J., & Solans, C. (2004). Formation B: Biointerfaces, 45, 66–75.
and stability of nano-emulsions. Advances in Colloid and Interface Vega, C., & Roos, Y. H. (2006). Invited review: Spray-dried dairy and
Science, 108–109, 303–318. dairy-like – Emulsions compositional considerations. Journal of Dairy
Thies, C. (2001). Microencapsulation: What it is and purpose. In P. Science, 89(2), 383–401.
Vilstrup (Ed.), Microencapsulation of food ingredients (pp. 1–29). Walstra, P. (2003). Physical chemistry of foods. New York, NY: Marcel
Surrey, England: Leatherhead Publishing. Dekker.