LWT - Food Science and Technology 131 (2020) 109519

Contents lists available at ScienceDirect

LWT - Food Science and Technology

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

Microencapsulation of pomegranate (Punica granatum L.) seed oil by T

complex coacervation: Development of a potential functional ingredient for
food application☆
André M.M. Costaa, Leticia K. Morettia, Grazieli Simõesb, Kelly A. Silvac, Verônica Caladod,
Renata V. Tonone,∗,1, Alexandre G. Torresa,1
a
Laboratório de Bioquímica Nutricional e de Alimentos, Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil
b
Departamento de Físico-Química, Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil
c
Departamento de Bromatologia (MBO), Universidade Federal Fluminense, Rio de Janeiro, Brazil
d
Escola de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil
e
Embrapa Agroindústria de Alimentos, Rio de Janeiro, Brazil

A R T I C LE I N FO A B S T R A C T

Keywords: The objective of this study was to develop a functional ingredient containing punicic acid, a bioactive conjugated
Pomegranate seed oil linolenic acid isomer, via microencapsulation of pomegranate seed oil (PSO) by complex coacervation. Whey
Punicic acid protein and gum Arabic were used as encapsulating agents and spray drying was applied as a hardening step. The
Complex coacervation effects of total polymer concentration (Cp) (2.2–7.8 g/100 mL) and wall material:oil ratio (WM:Oil) (0.5–5.0) on
Whey protein
microparticles physical-chemical characteristics (oil retention, microencapsulation efficiency (ME), punicic acid
Gum Arabic
content and particle size) were evaluated, according to a 22 rotatable central composite design. Microparticles'
morphology and surface composition were also assessed. Both Cp and WM:Oil ratio affected oil retention, ME,
punicic acid content and particle size. Intermediate values of Cp and WM:Oil ratio were considered the best
conditions for PSO encapsulation, resulting in the highest oil retention (near 100 g/100 g; and punicic acid
content, near 64 g/100 g fatty acid). High Cp and low WM:Oil ratios promoted microparticles’ agglomeration.
PSO microparticles rich in punicic acid were successfully produced by complex coacervation, enabling future use
of PSO as a functional ingredient in food products.

1. Introduction system and anti-diabetic properties, seem to be associated with its

singular fatty acid profile (Viladomiu, Hontecillas, Lu, & Bassaganya-
Pomegranate (Punica granatum) is a fruit appreciated for its pulp Riera, 2013). PSO is composed chiefly by conjugated linolenic acid
flavor and color, being mainly destined for the production of juice, isomers (cLnAs), specially punicic acid, which corresponds to ap-
nectar, jam and jellies (Shabbir et al., 2017). The seeds represent a proximately 70 g/100 g of total fatty acids in the oil (Fernandes et al.,
major pomegranate residue (3.7–7.9 g/100 g fruit weight) from juice 2015).
and jelly industries, consisting of a rich source of lipids (12–20 g/100 g) cLnA is a collective term for the positional and geometric isomers of
with an interesting chemical composition in terms of bioactive com- linolenic acid (C18:3), characterized by the presence of three con-
pounds (Fernandes et al., 2015), but they are largely discarded. Po- jugated double bounds, usually in positions Δ9,11,13, and Δ8,10,12 and
megranate seed oil (PSO) is known for its high contents of tocopherols, with varying combinations of geometrical configurations, cis or trans
phytosterols and phenolic compounds. However, the described func- (Cao, Gao, Chen, Chen, & Yang, 2006). Although cLnA isomers are
tional activities, such as cytotoxic effect, modulation of the immune emerging as potential bioactive nutrients, studies in humans evaluating

☆
Chemical compounds studied in this article:⟨-eleostearic acid (PubChem CID: 5281115); ®-eleostearic acid (PubChem CID: 5282820); catalpic acid (PubChem CID:
5385589); punicic acid (PubChem CID: 5281126).
∗
Corresponding author. Embrapa Agroindústria de Alimentos., Av. das Américas, 29501- Guaratiba, CEP 23020-470, Rio de Janeiro, RJ, Brazil.
E-mail addresses: andreme_1@hotmail.com (A.M.M. Costa), leticia_korin@hotmail.com (L.K. Moretti), grazielisimoes@iq.ufrj.br (G. Simões),
kalenkar@yahoo.com.br (K.A. Silva), calado@eq.ufrj.br (V. Calado), renata.tonon@embrapa.br, renata.tonon@embrapa.br (R.V. Tonon),
torres@iq.ufr.br (A.G. Torres).
1
These authors share senior authorship.

https://doi.org/10.1016/j.lwt.2020.109519
Received 10 December 2019; Received in revised form 22 March 2020; Accepted 29 April 2020
Available online 24 May 2020
0023-6438/ © 2020 Elsevier Ltd. All rights reserved.
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

Table 1
Oil retention, microencapsulation efficiency (ME), punicic acid content, moisture, water activity (aw), particle size, for the 11 trials of the experimental design.
(Mean ± SD).
F Cp Ratio Oil retention (g/100 g) ME Punicic acid content Moisture content (g/ aw Particle size
(g/100 mL) (WM:Oil) (g/100 g) (g/100 g fatty acid) 100 g)
D[0.5] (μm) Span

1 3 (−1) 1.15 (−1) 78.05 ± 3.23 50.00 ± 0.88 60.48 ± 0.54 3.81 ± 0.28 0.410 ± 0.011 8.43 ± 0.24 3.33 ± 0.08
2 3 (−1) 4.35 (+1) 75.00 ± 0.01 55.16 ± 0.11 60.49 ± 0.74 4.81 ± 0.28 0.329 ± 0.002 9.18 ± 1.53 2.24 ± 0.59
3 7 (+1) 1.15 (−1) 98.58 ± 1.06 43.17 ± 4.17 60.69 ± 1.83 1.91 ± 0.42 0.353 ± 0.001 10.08 ± 0.51 1.27 ± 0.17
4 7 (+1) 4.35 (+1) 88.64 ± 4.68 60.76 ± 2.09 59.95 ± 0.37 4.21 ± 0.01 0.288 ± 0.006 8.67 ± 0.25 1.29 ± 0.05
5 2.17 2.75 (0) 60.33 ± 2.71 61.40 ± 0.84 59.88 ± 0.78 5.30 ± 0.41 0.366 ± 0.006 8.36 ± 1.38 3.75 ± 0.12
(−1.41)
6 7.82 2.75 (0) 85.08 ± 1.11 62.92 ± 0.18 61.22 ± 0.19 3.21 ± 0.28 0.219 ± 0.006 10.96 ± 0.75 1.61 ± 0.21
(+1.41)
7 5 (0) 0.48 (−1.41) 106.19 ± 0.44 38.52 ± 1.92 62.12 ± 0.38 0.4 ± 0.28 0.285 ± 0.036 9.10 ± 0.10 1.53 ± 0.07
8 5 (0) 5.01 (+1.41) 83.23 ± 3.43 67.40 ± 5.31 61.08 ± 1.71 4.19 ± 0.34 0.230 ± 0.006 8.82 ± 0.44 0.99 ± 0.06
9a 5 (0) 2.75 (0) 95.2 ± 0.08 55.46 ± 4.62 62.93 ± 2.14 4.92 ± 0.71 0.283 ± 0.001 9.29 ± 0.09 1.54 ± 0.24
9a 5 (0) 2.75 (0) 94.28 ± 3.04 56.98 ± 0.59 64.22 ± 0.5 4.19 ± 0.52 0.379 ± 0.004 10.28 ± 0.12 1.41 ± 0.29
9a 5 (0) 2.75 (0) 105.41 ± 0.13 48.74 ± 0.47 63.85 ± 1.1 4.38 ± 0.28 0.331 ± 0.068 10.13 ± 1.08 1.92 ± 0.06

F: Formulation; acentral point; Cp: polymer concentration in the emulsion; WM: wall material aw: water activity; D[0.5]: maximum size (μm) of 50% analyzed
particles; span: particle size scattering index.

the health benefits of its acute or chronic consumption are still scarce, rotatable central composite design, aiming to define the best micro-
because these compounds, especially punicic acid, are restricted to PSO particle formulation. Oil retention, microencapsulation efficiency, pu-
and Trichosanthes kirilowii seed oil (Shabbir et al., 2017). The devel- nicic acid content and particle size were analyzed as responses.
opment of food products using PSO as a functional ingredient could Microparticles morphology and surface composition were also eval-
stimulate its consumption and help to clarify the oil bioactivity. How- uated.
ever, direct addition of PSO aiming at supplementing food products is
limited by its hydrophobic nature and by the high susceptibility of cLnA 2. Material and methods
isomers to lipid oxidation when exposed to oxygen and light (Yang,
Cao, Chen, & Chen, 2009). Thus, these functional lipids should be 2.1. Materials
protected in order to preserve their physical and chemical stability,
avoiding oxidative rancidity and nutritional losses. Cold-pressed commercial pomegranate seed oil (PSO)
Microencapsulation is a “packing” technique in which an active (C16:0 = 2.25 g/100 g, C18:0 = 1.86 g/100 g, C18:1n-9 = 4.52 g/
ingredient is covered by a wall material, being often used to protect 100 g, C18:1n-7 = 0.41 g/100 g, C18:2n-6 = 5.29 g/100 g,
unstable molecules from the interaction with other components and the C20:0 = 0.61 g/100 g, C20:1 n-9 = 0.78 g/100 g, Total
adjacent environment during food processing and storage (Gouin, cLnA = 84.3 g/100 g) (Oneva Food Co®, Istanbul, Turkey) was used as
2004). core material. Whey protein isolate (WPI) (Alibra®, São Paulo, Brazil)
Complex coacervation is an encapsulation method that consists of a and Gum Arabic (GA) (Instantgum BA®, Colloides Naturels, São Paulo,
liquid-liquid phase separation phenomenon that occurs when electro- Brazil) were used as wall materials for particles production.
statically opposite charged biopolymers are subjected to specific con- A commercial mixture of fatty acid methyl esters (37-component
ditions, producing aggregates (coacervates) that promptly deposit on FAME mix; Supelco, Bellefonte, PA, US) and individual cLnA isomers
the oil droplets. Compared to other encapsulation techniques, complex (punic acid, alpha-eleostearic acid, catalpic acid and beta-eleostearic
coacervation is able to produce microparticles with higher micro- acid; Larodan AB, Solna, Sweden) were used as standards for fatty acid
encapsulation efficiency, using high core load and low wall material identification in gas-chromatographic analyzes. All solvents used were
concentration (Gouin, 2004). The process performance and the physi- HPLC grade from Tedia (São Paulo, Brazil) and all reagents used were
cochemical properties of particles produced by complex coacervation from Merck (Darmstadt, Germany).
can be influenced by parameters such as: total polymers concentration,
protein:polyssacharide ratio, core:wall material ratio, pH, salt con- 2.2. Production of the PSO's microparticles by complex coacervation
centration and others (Weinbreck, de Vries, Schrooyen, & de Kruif,
2003). For established complex coacervate systems, such as whey PSO's microparticles were produced as follows: firstly, the wall
protein:gum Arabic and gelatin:gum Arabic, some factors (protein:po- materials (WPI and GA) were separately weighed and dissolved in
lyssacharide ratio, pH and salt concentration) are well known distilled water under magnetic stirring during 30 min, at room tem-
(Weinbreck et al., 2003; Weinbrreck, Tromp, & de Kruif, 2004), while perature, to obtain the solutions with Cp values described in Table 1.
other parameters need to be studied in order to provide satisfactory The solutions' pH was adjusted to 7.0 with HCl (0.5 mol/L) and NaOH
process performance, when new core nutrients are used. Whey protein- (0.5 mol/L), and the WPI:GA ratio was 2:1 (Weinbreckt et al., 2004).
gum Arabic system has been successfully used to encapsulate a wide After dissolution, PSO was added dropwise to the WPI solution under
range of hydrophobic materials (Eratte et al., 2015; Eratte, Wang, continuous stirring (16,000 rpm) during 5 min with an Ultra-Turrax
Dowling, Barrow, & Adhikari, 2014; Weinbreckt, Minorf, & Kruif, homogenizer (T25-IKA®, IKA, Wilmington, US), to produce a stable
2004). Nevertheless, to the best of our knowledge, microencapsulation emulsion. Then, GA solution was mixed with the previous emulsion and
of PSO by complex coacervation using this system has not yet been homogenized for 1 min at 16,000 rpm. Finally, emulsion's pH was ad-
studied. justed to 3.75 by adding HCl (1.0 mol/L) (Eratte et al., 2014) aiming at
This work aimed at investigating complex coacervation as a suitable inducing electrostatic interactions between WPI and GA, thus forming
method for microencapsulation of pomegranate seed oil, using whey the wet microparticles. Microencapsulation process was carried out at
protein and gum Arabic as encapsulating system. The effect of total 25 °C, for 10 min.
polymer concentration (Cp) and wall material:oil ratio (WM:Oil ratio) A rotatable central composite design was applied to evaluate the
on the dried microparticles properties was assessed, according to a effect of total polymer concentration (Cp) (2.2–7.8 g/100 mL) and wall

2
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

material:oil ratio (WM:Oil ratio) (0.5–5.0) on the microencapsulation of fatty acid.

PSO by complex coacervation, according to Table 1.
The following polynomial equation was fitted to data (Equation 2.5.4. Moisture content and water activity (aw)
(1)): Moisture content was determined in all samples using a moisture
balance with infrared radiation heating (MA35 Mettler Toledo, Urdorf,
y = β0 + β1 x1 + β2 x2 + β11 x12 + β22 x 22 + β12 x1 x2 (1) Switzerland). Water activity was assessed in a water activity analyzer
where βn are constant regression coefficients, y is the response for dried (LabMaster-aw, Novasina AG, Lanchen, Switzerland).
particles physicochemical properties (oil retention, microencapsulation
efficiency, punicic acid content, moisture content, water activity or 2.5.5. Particle size
particle size), and x1 and x2 are the coded independent variables (Cp PSO’ microparticles were dispersed in isopropanol and analyzed in a
and WM:Oil ratio). particle size analyzer SDC-Microtrac S3500 (Microtrac,
Montgomeryville, PA, US) by the scattering pattern of a transverse laser
2.3. Drying of the PSO's microparticles light. Results were reported as D[0.5] and scattering index (Span)
(Equation (4)), which are defined as the mean diameter (μm) and
Immediately after preparation, the wet microparticles were dried in width distribution of particles size range, respectively.
a laboratory scale spray dryer (SD-06AG, Lab Plant, North Yorkshire, D [0.9] − D [0.1]
Span =
UK) equipped with a 0.5 mm nozzle. The drying conditions were: inlet D [0.5] (4)
air temperature of 180 °C, outlet air temperature of 66 ± 4 °C, air flow
pressure of 2.5 × 105 Pa and feed flow rate of 0.5 L/h. Dried micro- where: D[0.9], D[0.1] and D[0.5] are the diameters at 10, 50, and 90%
particles were collected, sealed in plastic bags under vacuum, covered cumulative volume, respectively.
in aluminum foil and stored at −80 °C for further characterization.
2.5.6. Morphology
PSO microparticles’ morphology was evaluated by scanning electron
2.4. Wet microparticles morphology
microscopy (SEM). Samples were directly deposited on carbon con-
ductive tape on aluminum SEM stubs, and coated with a thin gold layer,
Wet microparticles morphology was evaluated in an optical micro-
using a gold-sputter (Desk V, Denton Vacuum®, Moorestown, NJ, US).
scope (Eclipse E200; Nikon®, Tokyo, Japan) coupled to a digital camera
The samples were analyzed using a Tescan Vega 3 SEM (Tescan®,
(Evolution VF, Media Cybernetics, Rockville, MD, US). Prior to ob-
Kohoutovice, Czech Republic) operated at 15 kV.
servation in the microscope, microparticles were collected from the
formulations by centrifugation (11,739 ⋅g, 10 min) and were diluted
2.5.7. X-ray photoelectron spectroscopy (XPS) of microparticles
twice with ultrapure water acidified with 0.5 mol/L HCl (pH 3.7).
XPS analysis was performed on a UHV Xi ESCALAB 250 (Thermo
Images were registered at 400 × and 1000 × magnifications.
Fisher Scientifics, US) spectrometer equipped with a hemispherical
electron energy analyzer. The XPS spectra were collected using mono-
2.5. Dried microparticles analysis
chromatic Al Kα X-ray source (1486.6 eV) and an electron emission
angle of 90° with the surface. Survey scans were recorded with 1 eV
2.5.1. Oil retention
steps and 100 eV analyzer pass energy. The high-resolution C1s spectra
Total oil content present in the microparticles was determined in a
were recorded with 0.1 eV steps and 25 eV pass energy analyzer. The
Soxhlet apparatus (SER 148; Velp® Scientifica, Usmate, Italy), ac-
linearity of the energy scale was checked using Cu (932.7 eV), Ag
cording to Goula and Adamopoulos (2012), with few modifications. The
(368.3 eV) and Au (84.0 eV) lines. Data processing was performed using
oil extracted with petroleum ether added with BHT was weighed, re-
Thermo Avantage software. Peak fitting was carried out with
suspended with a known volume of hexane with BHT (0.05 g/100 mL)
Lorentzian/Gaussian ratio of 30%/70%.
and stored at −20 °C until further analysis.
Oil retention was defined as the percentage of total oil in the final
2.6. Statistical analysis
powder to that initially added in the feed emulsion and was calculated
as follows (Equation (2)):
For the central composite design, analysis of variance (ANOVA), test
Mass of total oil (g ) of lack of fit, determination of regression coefficients and the con-
Oil retention (g / 100 g ) = × 100
Mass of oil in the feed emulsion (g ) struction of response surface (3D) graphs, Statistica 7.0 (StatSoft, Tulsa,
(2) US) software was used. Statistical comparisons were based on at least
triplicate results, and all data were presented as mean and standard
deviation. The influence of the encapsulation process on cLnA isomer
2.5.2. Microencapsulation efficiency (ME) distribution was evaluated by comparing a mean obtained from the
Surface oil was determined according to Sankarikutty, Sreekumar, distribution of each cLnA isomers from the 11 formulations tested, with
Narayanan, and Mathew (1988), with the modifications proposed by a non-encapsulated sample, using t-test. Data were analyzed with
Kouassi et al. (2012). ME was calculated according to Equation (3): GraphPad Prism v.6.0 (GraphPad software 2012, La Jolla, CA, US), and
Total oil (g ) − Surface oil (g ) the bar graphs displayed in this work were also plotted using this
ME (g / 100 g ) = × 100 program. Significance level was established at p < 0.05.
Total oil (g ) (3)

3. Results and discussion

2.5.3. Punicic acid content
PSO fatty acids extracted from the microparticles were methylated 3.1. Morphology of wet microparticles
by a base-catalyzed transesterification procedure (Kramer et al., 1997)
to avoid isomerization. PSOs’ fatty acid methyl esters (FAME) were A noticeable coacervate layer around the oil droplets was noticeable
analyzed by gas chromatography with a flame ionization detector (GC- in most of the formulations (Fig. 1), indicating the effectiveness of the
FID) and equipped with a split/splitless injector (GC-2010 chromato- encapsulation process. Formulations with higher oil content showed
graph; Shimadzu, Japan) exactly as described by Costa, Silva, and larger oil droplets (Fig. 1a, c and 1g), which can suggest that coales-
Torres (2019). Fatty acid composition results were expressed as g/100 g cence may be starting after homogenization ceased, indicating that

3
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

Fig. 1. Wet capsules' micrographs of the experimental design formulations obtained by optical microscopy: a) 1, Cp: 3 g/100 mL; WM:Oil: 1.15; b) 2, Cp: 3 g/100 mL;
WM:Oil: 4.35; c) 3, Cp: 7 g/100 mL; WM:Oil: 1.15; d) 4, Cp: 7 g/100 mL; WM:Oil: 1.15; e) 5, Cp: 2.17 g/100 mL; WM:Oil: 2.75; f) 6, Cp: 7.82 g/100 mL; WM:Oil: 2.75;
g) 7, Cp: 5 g/100 mL; WM:Oil: 0.48; h) 8, Cp: 5 g/100 mL; WM:Oil: 5.01; i) 9, Cp: 5 g/100 mL; WM:Oil: 2.75; j) Coacervate (C); j2) Coacervate (C); l) Formulation 9
produced at pH 7. Wet capsules were prepared in pH 3.7, whey protein: gum Arabic ratio (2:1) and reaction time of 10 min a, b, c, d, e, f, g, h, i, j and m:
magnification of 400 × ; j2: magnification of 1000 × .

these structures were less stable than the other formulations. For- between the core and microcapsule's surface should be kept in order to
mulation 9 was used as a negative control at pH = 7.0, in which WPI avoid oil droplets migration to particles' external layer, thus reducing
and gum Arabic have similar net charges and thus suppressing coa- microparticles surface oil.
cervation. As expected, there was no coacervate layer adsorbed on the
oil droplets in this formulation (Fig. 1l). The coacervates without PSO 3.2. Dried microparticles
showed a transparent gel-like structure (Fig. 1j and 1j2) (Weinbreckt
et al., 2004). All formulations showed multi-cored microparticles PSO's microparticles moisture content and aw varied from 0.4 to
(Fig. 1), which generally occurs when high homogenization rates are 5.3 g/100 g and 0.219–0.410, respectively, which are values commonly
used during the emulsification step prior to coacervation (Eratte et al., observed in microparticles produced by spray drying (moisture ≤ 6 g/
2014; Kaushik, Dowling, McKnight, Barrow, & Adhikari, 2016; Liu, 100 g and aw ≤ 0.6) (Klaypradit & Huang, 2008; Reineccius, 2004).
Low, & Nickerson, 2010).
Polymer concentration did not show a clear effect on particles
3.2.1. Experimental design
morphology, while the WM:Oil ratio influenced the coacervate layer
The results for each response analyzed in the central composite
thickness. When comparing formulations with the same polymer con-
design are shown in Table 1. The regression coefficients for the poly-
centration and distinct WM:Oil ratios (F1 vs. F2; F3 vs. F4; F7 vs. F8 vs.
nomial equation, F values and determination coefficients (R2) for each
F9; Fig. 1a vs 1b; Fig. 1c vs 1d and Fig. 1g vs 1h vs 1i), it is possible to
response are shown in Table 2. The calculated F values were higher
notice that an increase in the WM:Oil ratio resulted in a thicker coa-
than the tabulated ones for all the evaluated responses, except for ME,
cervate layer. This same trend was also observed by Ma, Zhao, Wang,
indicating that this response could not be predicted by the adjusted
and Sun (2019), using the same technology to encapsulate methyl
model.
oleate applying gelatin:Gum Arabic system. According to Goula and
Adamopoulos (2012), microparticles with a thicker coacervate layer are
more suitable to the atomization process, because a minimum length 3.2.1.1. Oil retention. The presence of unbound oil on the dryer wall
was only observed in formulation 7, which had the highest oil content

4
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

Table 2 higher than those previously observed in the microencapsulation of

Coded second-order regression coefficients, F values and determination coeffi- flaxseed and tuna oil by complex coacervation (Eratte et al., 2014;
cients (R2) for oil yield, microencapsulation efficiency (ME), punicic acid con- Kaushik et al., 2016; Liu et al., 2010). Oil retention was positively
tent and particle size scattering index (Span). influenced by Cp (Table 2), until a maximum Cp value ranging from 5 to
Coefficient Oil retention (g/ ME Punicic acid Span 7 g/100 mL (Fig. 2a). Previous studies suggested that the increase of Cp
100 g) (g/100 g) content implies in an increase of emulsion viscosity (Frascareli, Silva, Tonon, &
(g/100 g fatty acid) Hubinger, 2011; Goula & Adamopoulos, 2012; Sahin-Nadeem & Özen,
β0 98.29** 53.75** 63.67** 1.62**
2014). Viscosity of the coacervate layer is important to form a resistant
β1 8.66* 0.12 0.19 −0.76** barrier around the oil droplet, thus favoring oil retention during the
β11 −12.51** 2.90 −1.73** 0.55** drying process.
β2 −5.69 7.96* −0.27 −0.23
β22 −1.44 −1.73 −1.21* −0.17
β12 −1.72 3.11 −0.19 0.28 3.2.1.2. Microencapsulation efficiency. Microencapsulation efficiency
Fcalculated 13.91 4.6 8.4 57.98 (ME) varied from 38.5 g/100 g–67.4 g/100 g (Table 1) and was
Ftabulated 5.05 5.05 5.05 5.05
R2 0.932 0.821 0.896 0.980
positively influenced by the WM:Oil ratio (Table 2), thus formulations
with WM:Oil ratios equal to or higher than 2.75 showed the highest
β0: mean; β1: Total polymer concentration linear; β11: Total polymer con- values of ME (Table 1). For this variable, a response surface was not
centration quadratic; β2: Wall material: oil (WM:Oil) ratio linear; β22: WM:Oil presented, due to the lack of fit of the proposed model (R2 = 0.82 and
ratio quadratic; β12: Total polymer concentration × WM:Oil ratio. *significant Fcalculated < Ftabulated) (Table 2).
at p < 0.06; **significant at p < 0.05. The influence of WM:Oil ratio on ME observed in this study has also
been described in previous works (Jun-xia, Hai-yan, & Jian, 2011;
(lowest WM:Oil ratio) and showed the lowest oil retention Kaushik et al., 2016; Yang, Gao, Hu, Li, & Sun, 2015). Similarly to the
(60.3 ± 2.71 g/100 g), while all other formulations showed oil oil retention, this result could be associated to the higher proportions of
retention between 75 and 106 g/100 g. These values were sensibly oil droplets close to the drying surface, which shortens the diffusion

Fig. 2. Response surfaces of: a) oil yield; b) punicic acid content; and c) Span.
*Cp: Total polymer concentration, WM = wall material.

5
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

any independent variable, and an overall narrow particle size

distribution, demonstrated by the low Span values (Table 1).
According to the experimental design results, only Cp significantly
affected the Span value of PSO's microparticles (Table 2), and low span
values were found in microparticles with Cp between 5 and 7.82 g/
100 mL (Fig. 2c). Particles with narrow particle size distribution are
appealing for food application, as this characteristic is considered im-
portant to promote a sustained core release (Walton & Mumford, 1999).
As the coacervates were not crosslinked, microparticles' structure
could have been partially broken during atomization. Thus, it is pos-
sible that the particles’ diameters observed refer to particle fragments,
and not to their original structure. This hypothesis could be evaluated
by comparing crosslinked and non-crosslinked structures after atomi-
zation.

3.2.1.5. Dried PSO microparticles’ morphology. Surface morphology of

the PSO microparticles is shown in Fig. 4. Most of the formulations
showed spherical shape and variable sizes with numerous
invaginations, which is typical of spray-dried microparticles (Ré,
Fig. 3. Fatty acid composition in pomegranate seed oil and in microparticles 1998). Similar morphology was observed by Eratte et al. (2015;
produced with different formulations. The encapsulation process influenced all 2014) and Rutz et al. (2017), which also microencapsulated edible
cLnA isomers distribution (t-test, p < 0.05). * : cLnA 4; : cLnA 3; : cLnA 2; oils by complex coacervation and used spray drying as a hardening/
: cLnA 1; : Beta-eleostearic acid; : Catalpic acid; : Alfa-eleostearic acid; recovery step. However, as mentioned, the original structure of the non-
Punicic acid. crosslinked coacervates could have been broken during atomization,
resulting in a new arrange of polymer material with the same
path length to the drying microcapsule external layer. Wet micro- morphology of traditional spray-dried particles. Thus, we cannot rule
particles’ micrographs in Fig. 1 agree with this trend, as it is possible to out that the particles observed in Fig. 4 are in fact not the original
notice that WM:Oil ratio influenced the coacervate layer thickness. coacervates.
Additionally, at low WM:Oil ratios (high oil loads), the Cp could have In the formulations with a WM:Oil ratio of 2.75 (Fig. 3e, f and 3i),
not been enough to completely cover the oil droplets and prevent oil the increase of Cp above 5 g/100 mL promoted noticeable aggregation
droplets coalescence. Consequently, large droplets were surrounded by problems (Fig. 4). This phenomenon could be associated to an over-
a thin coacervate layer, while small droplets were encapsulated by a balance of the wall materials, which intensified microparticles adhe-
matrix of coacervate, as shown in Fig. 1. sion, due to an increase in the feed emulsion viscosity (Yang et al.,
2015). Regarding oil load, formulations with a WM:Oil ratio of 1.15 or
3.2.1.3. Punicic acid retention in microparticles. PSO is a rich source of a lower also showed aggregation problems. These problems were more
family of bioactive fatty acids, namely cLnA (Fig. 3). GC analysis evident in Formulation 7 (Fig. 4g), where microparticles’ structure
showed eight cLnA isomers, from which four (punicic, alpha- collapsed, and it was not possible to observe surface features of in-
eleostearic, catalpic and beta-eleostearic acids) were identified based dividual microparticles. This could be attributed to the high surface oil
on comparisons with standards’ retention times and mass spectrometric content present in these formulations, which may have enhanced mi-
fragmentation profiles (Supplementary Fig. 1; Costa et al., 2019). croparticles stickiness (Goula & Adamopoulos, 2012). Interesting mor-
Punicic acid is the major cLnA in PSO and its content is associated to phological characteristics for food application were demonstrated in
the oil's functional properties (Viladomiu et al., 2013). Being PSO WM:Oil ratios between 2.75 and 4.35, such as spherical shape micro-
highly oxidizable because of its high cLnA content, the encapsulation particles with minimal structure break.
process promoted a small but significant loss of punicic acid (Fig. 3),
possibly by oxidation. In parallel, beta-eleostearic acid and other minor 3.3. Selection of the best formulation
cLnA isomers (cLnA 1, cLnA 2, cLnA 3 and cLnA 4; Supplementary
Fig. 1) contents increased from 1.5- to 3.6-fold, indicating that part of The variables used to select the best formulation were oil retention,
punicic acid was isomerized, probably because of the high temperatures ME and punicic acid content. Based on these responses, the inter-
applied in the drying process (Sahin-Nadeem & Afşin Özen, 2014) and mediate values of Cp and WM:Oil ratio (Cp = 5 g/100 mL, WM:Oil
because the trans-isomers are more stable than the cis-isomers of cLnA ratio = 2.75) were chosen. Although this formulation did not show the
(Giua, Blasi, Simonetti, & Cossignani, 2013). highest values of ME, it was located in the region near the optimal
Punicic acid content varied only slightly among the PSO micro- values of oil retention and punicic acid content (Fig. 2a and b), showing
particles formulations (from 59.9 to 64.2 g/100 g oil, Table 1). How- the highest content of the featured bioactive compound in PSO and a
ever, these marginal differences might result in a final product with considerable load of PSO per gram of dried microparticles (26.6 g PSO/
sensibly different punicic acid content, as microparticles' oil load varied 100 g microparticles).
up to 4-fold among formulations, for instance, 5.01 WM:Oil ratio had
16.6 g PSO/100 g of microparticles vs. 0.48 WM:Oil ratio that had 3.4. XPS analysis
67.7 g PSO/100 g of microparticles. The response surface graph shows a
region of maximum punicic acid content on the intermediate levels of The concentration of different carbon species was determined by
Cp and WM:Oil ratio (Fig. 2b). ME values were positively associated to curve fitting the high-resolution C1s spectra. The spectra of the WM
punicic contents, consequently, having low amounts of non-en- (coacervate without oil), WPI, GA, the elected formulation (F9) and F7
capsulated punicic acid on the particle surface might delay PSO oxi- (formulation with the lowest ME: 38.5 g/100 g) showed carbon com-
dation. ponents with the following assignments: C1 (C–C, C–H); C2 (C–O, C–N);
C3 (C]O, O–C–O, N–C]O) and C4 (O–C]O). The C1s high resolution
3.2.1.4. Particle size. PSO microparticles showed a small variation in spectra are shown in Fig. 5b, c, 5d, 5e and 5f. The XPS analysis of oil
the mean diameter (8.36–10.96 μm), with no significant influence of was impractical in ultra-high vacuum as it is liquid.

6
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

Fig. 4. Dried microparticles micrographs of the experimental design formulations obtained by scanning electron microscopy (SEM): a) 1, Cp: 3 g/100 mL; WM:Oil:
1.15; b) 2, Cp: 3 g/100 mL; WM:Oil: 4.35; c) 3, Cp: 7 g/100 mL; WM:Oil: 1.15; d) 4, Cp: 7 g/100 mL; WM:Oil: 1.15; e) 5, Cp: 2.17 g/100 mL; WM:Oil: 2.75; f) 6, Cp:
7.82 g/100 mL; WM:Oil: 2.75; g) 7, Cp: 5 g/100 mL; WM:Oil: 0.48; h) 8, Cp: 5 g/100 mL; WM:Oil: 5.01; i) 9, Cp: 5 g/100 mL; WM:Oil: 2.75. Bar = 20 μm. White
arrows indicate cracks and open pores.

The C1s spectrum of WPI is characterized by a particularly strong encapsulated PSO showed higher contributions of C4 than the control
C1 contribution (Fig. 5c). This characteristic is not expected for a with no PSO (WPI, GA, WM; Table 3). This result could also be extra-
protein spectrum but can result from the presence of lipids on the polated to all formulations (Fig. 6), because C4 peek intensity was ne-
surface. The GA peaks (Fig. 5d) showed, as expected, a typical poly- gatively influenced by ME results (Table 1), indicating agreement be-
saccharide spectrum, dominated by the C–O peak. This was confirmed tween XPS results and the washing method described earlier. Jafari,
in the survey spectrum by a strong oxygen signal and very little or no Assadpoor, Bhandari, and He (2008) microencapsulated fish oil by
nitrogen signal (Fig. 5a). Surface composition of coacervate WM pre- spray drying and also observed an agreement between both methods.
sented characteristics of both WPI and GA (Fig. 5b). F9 and F7 spectra Additionally, XPS analysis was also capable to indicate interactions
combine characteristics of WPI spectrum, such as prominent C1 che- between WPI and GA, as the C3 and C4 chemical environment in-
mical environment (Fig. 5e and f) and the presence of nitrogen in the creased in WM, F9 and F7 (Table 3), probably because a higher ex-
survey spectra (Fig. 5a), indicating that different concentrations of Cp posure of WPI (NH3+) and GA (COO−) interaction sites on micro-
and WM:Oil ratio results in variations in microparticles surface com- particles surface, resulting from pH adjustment (Eratte et al., 2015).
position. XPS analysis was tentatively used to evaluate microparticles
surface composition, because this method does not modify samples’
4. Conclusion
surface composition, as the classical washing methods applied earlier.
Therefore, after evaluating the C1-4 chemical environment, a relation
PSO was successfully microencapsulated by complex coacervation,
between the O–C]O peak was observed, highlighted as acid/ester
showing a minimal impact on PSO isomer distribution. The feed
groups, and a higher content of surface oil, as samples with
emulsion formulation had a significant influence on the coacervation

7
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

Fig. 5. Microparticles' X-ray photoelectron spectroscopy (XPS) analysis. a) XPS survey spectra of wall material (WM), whey protein isolate (WPI), gum Arabic (GA),
elected formulation (F9: Cp: 5 g/100 mL; WM:Oil: 2.75), and formulation with the lowest microencapsulation efficiency (F7: Cp: 5 g/100 mL WM:Oil: 0.48); and XPS
C1s high-resolution spectra of b) WM, c) WPI, d) GA, e) F9, f) F7.
*C1 (C–C, C–H); C2 (C–O, C–N); C3 (C]O, O–C–O, N–C]O) and C4 (O–C]O). a) : F7; : F9; : WM; : GA; : WPI. In b), c), d), e) and f), : C1s data; : C1; : C2;
: C3; : C4; : background; : fit.

Table 3 process parameters manipulation enabled the design of PSO's micro-

Surface elemental composition of formulations (F9: Cp: 5 g/100 mL; WM:Oil: particles with physical-chemical properties that allows its application in
2.75 and F7: Cp: 5 g/100 mL; WM:Oil: 0.48) and components (WM, GA, WPI), food products. However, it remains to be confirmed if the original
determined by X-ray photoelectron spectroscopy (XPS). coacervates' structure was kept during spray drying, since the high
Elemental Composition WM GA WPI F9 F7 pressure and turbulent flow inside the atomizer could have broken the
(%) (%) (%) (%) (%) particles structure without altering the gross morphology of the ob-
tained powder.
C1 (C–C, C–H) 52.1 16.4 74.3 71.7 74.3
C2 (C–O, C–N) 21.9 77.3 19.2 19.0 16.5
C3 (C]O, O–C–O, N–C]O) 22.6 4.5 4.6 5.0 0.4
CRediT authorship contribution statement
C4 (O–C]O) 3.4 1.8 1.9 4.3 8.8

André M.M. Costa: Conceptualization, Methodology, Formal ana-

process and microparticles physical-chemical and morphological lysis, Investigation, Data curation, Writing - original draft,
properties. Among the physical-chemical properties evaluated in this Visualization, Project administration. Leticia K. Moretti: Investigation,
study, oil retention, ME, punicic acid content and Span values were Visualization. Grazieli Simões: Methodology, Formal analysis,
influenced by experimental design factors. Moreover, surface analysis Investigation, Visualization, Writing - original draft. Kelly A. Silva:
determined by XPS was accurate, being especially indicated for porous Methodology, Formal analysis, Investigation, Visualization, Writing -
wall systems aiming at avoiding extraction of the encapsulated oil. original draft. Verônica Calado: Methodology, Formal analysis.
Microparticles produced with 5 g/100 g of Cp and WM:Oil ratio of 2.75 Renata V. Tonon: Conceptualization, Methodology, Resources, Data
were selected as the most promising ones, due to the highest punicic curation, Writing - review & editing, Supervision, Funding acquisition.
acid content and oil retention. In the present study, coacervation Alexandre G. Torres: Conceptualization, Methodology, Resources,
Data curation, Writing - review & editing, Supervision, Funding

8
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

Fig. 6. Microparticles' X-ray photoelectron spectroscopy (XPS) analysis. XPS C1s high-resolution spectra of Formulations: a) 1, Cp: 3 g/100 mL; WM:Oil: 1.15; b) 2,
Cp: 3 g/100 mL; WM:Oil: 4.35; c) 3, Cp: 7 g/100 mL; WM:Oil: 1.15; d) 4, Cp: 7 g/100 mL; WM:Oil: 1.15; e) 5, Cp: 2.17 g/100 mL; WM:Oil: 2.75; f) 6, Cp: 7.82 g/
100 mL; WM:Oil: 2.75; g) 8, Cp: 5 g/100 mL; WM:Oil: 5.01) and h) XPS survey spectra of Formulations 1, 2, 3, 4, 5, 6 and 8.
*In a), b), c), d), e), f) and g), : C1s data; : C1; : C2; : C3; : C4; : background; : fit; and in h) : F1; : F2; : F3; : F4; : F5; : F6; : F8.

9
A.M.M. Costa, et al. LWT - Food Science and Technology 131 (2020) 109519

acquisition. 639–652.
I Ré, M. (1998). Microencapsulation by spray drying. Drying Technology, 16, 1195–1236.
Jafari, S. M., Assadpoor, E., Bhandari, B., & He, Y. (2008). Nano-particle encapsulation of
Declaration of competing interest fish oil by spray drying. Food Research International, 41, 172–183.
Jun-xia, X., Hai-yan, Y., & Jian, Y. (2011). Microencapsulation of sweet orange oil by
complex coacervation with soybean protein isolate/gum Arabic. Food Chemistry, 125,
The authors declare that there are no conflicts of interest. 1267–1272.
Kaushik, P., Dowling, K., McKnight, S., Barrow, C. J., & Adhikari, B. (2016).
Acknowledgements Microencapsulation of flaxseed oil in flaxseed protein and flaxseed gum complex
coacervates. Food Research International, 86, 1–8.
Klaypradit, W., & Huang, Y. W. (2008). Fish oil encapsulation with chitosan using ul-
This work was supported by CAPES (Finance code 001), CNPq trasonic atomizer. LWT - Food Science and Technology, 41(6), 1133–1139.
(grants number 432484/2016-7, 309558/2015-8 and 310659/2018-3) Kouassi, G. K., Teriveedhi, V. K., Milby, C. L., Ahmad, T., Boley, M. S., Gowda, N. M., et al.
(2012). Nano-microencapsulation and controlled release of linoleic acid in biopo-
and FAPERJ (grant numbers E-26/010.001277/2015, E-26/203.197/
lymer matrices: Effects of the physical state, water activity, and quercetin on oxi-
2015 and E-26/203.294/2016) (Brazil). AGT and RVT are recipients of dative stability. Journal of Encapsulation and Adsorption Sciences, 2, 1–10.
CNPq and FAPERJ scholarships, AMMC was a recipient of a CNPq PhD Kramer, J. G., Fellner, V., Dugan, M. R., Sauer, F., Mossoba, M., & Yurawecz, M. (1997).
studentship. We are also grateful to Brenda Duarte Gralha for assistance Evaluating acid and base catalysts in the methylation of milk and rumen fatty acids
with special emphasis on conjugated dienes and total trans fatty acids. Lipids, 32,
in some of the laboratory analysis. 1219–1228.
Liu, S., Low, N. H., & Nickerson, M. T. (2010). Entrapment of flaxseed oil within gelatin-
Appendix ASupplementary data gum Arabic capsules. JAOCS. Journal of the American Oil Chemists’ Society, 87,
809–815.
Ma, T., Zhao, H., Wang, J., & Sun, B. (2019). Effect of processing conditions on the
Supplementary data to this article can be found online at https:// morphology and oxidative stability of lipid microcapsules during complex coa-
doi.org/10.1016/j.lwt.2020.109519. cervation. Food Hydrocolloids, 87, 637–643.
Reineccius, G. A. (2004). The spray drying of food flavors. Drying Technology, 22(6),
1289–1324 In press.
References Rutz, J. K., Borges, C. D., Zambiazi, R. C., Crizel-Cardozo, M. M., Kuck, L. S., & Noreña, C.
P. Z. (2017). Microencapsulation of palm oil by complex coacervation for application
in food systems. Food Chemistry, 220, 59–66.
Cao, Y., Gao, H. L., Chen, J. N., Chen, Z. Y., & Yang, L. (2006). Identification and char-
Sahin-Nadeem, H., & Özen, M. A. (2014). Physical properties and fatty acid composition
acterization of conjugated linolenic acid isomers by Ag+-HPLC and NMR. Journal of
of pomegranate seed oil microcapsules prepared by using starch derivatives/whey
Agricultural and Food Chemistry, 54, 9004–9009.
protein blends. European Journal of Lipid Science and Technology, 116, 847–856.
Costa, A. M. M., Silva, L. O., & Torres, A. G. (2019). Chemical composition of commercial
Sankarikutty, B., Sreekumar, M. M., Narayanan, C. S., & Mathew, A. G. (1988). Studies on
cold-pressed pomegranate (Punica granatum) seed oil from Turkey and Israel, and the
microencapsulation of cardamom oil by spray drying technique. Journal of Food
use of bioactive compounds for samples' origin preliminary discrimination. Journal of
Science & Technology, 25, 325–355.
Food Composition and Analysis, 75, 8–16.
Shabbir, M. A., Khan, M. R., Saeed, M., Pasha, I., Khalil, A. A., & Siraj, N. (2017). Punicic
Eratte, D., McKnight, S., Gengenbach, T. R., Dowling, K., Barrow, C. J., & Adhikari, B. P.
acid: A striking health substance to combat metabolic syndromes in humans. Lipids in
(2015). Co-encapsulation and characterisation of omega-3 fatty acids and probiotic
Health and Disease, 16, 1–9.
bacteria in whey protein isolate-gum Arabic complex coacervates. Journal of
Viladomiu, M., Hontecillas, R., Lu, P., & Bassaganya-Riera, J. (2013). Preventive and
Functional Foods, 19, 882–892.
prophylactic mechanisms of action of pomegranate bioactive constituents. Evidence-
Eratte, D., Wang, B., Dowling, K., Barrow, C. J., & Adhikari, B. P. (2014). Complex
based Complementary and Alternative Medicine, 2013.
coacervation with whey protein isolate and gum Arabic for the microencapsulation of
Walton, D., & Mumford, C. (1999). The morphology of spray-dried particles: The effect of
omega-3 rich tuna oil. Food Funct. 5, 2743–2750.
process variables upon the morphology of spray-dried particles. Chemical Engineering
Fernandes, L., Pereira, J. A., Lopéz-Cortés, I., Salazar, D. M., Ramalhosa, E., & Casal, S.
Research and Design, 77, 442–460.
(2015). Fatty acid, vitamin E and sterols composition of seed oils from nine different
Weinbreck, F., de Vries, R., Schrooyen, P., & de Kruif, C. G. (2003). Complex coacervation
pomegranate (Punica granatum L.) cultivars grown in Spain. Journal of Food
of whey proteins and gum Arabic. Biomacromolecules, 4, 293–303.
Composition and Analysis, 39, 13–22.
Weinbreckt, F., Minorf, M., & Kruif, C. G. (2004). Microencapsulation of oils using whey
Frascareli, E. C., Silva, V. M., Tonon, R. V., & Hubinger, M. D. (2011). Food and
protein/gum Arabic coacervates. Journal of Microencapsulation, 21, 667–679.
Bioproducts Processing Effect of process conditions on the microencapsulation of
Weinbrreck, F., Tromp, R. H., de Kruif, & De Kruif, C. G. (2004). Composition and
coffee oil by spray drying. Food and Bioproducts Processing, 90, 413–424.
structure of whey protein/gum Arabic coacervates. Biomacromolecules, 5, 1437–1445.
Giua, L., Blasi, F., Simonetti, M. S., & Cossignani, L. (2013). Oxidative modifications of
Yang, L., Cao, Y., Chen, J. N., & Chen, Z. Y. (2009). Oxidative stability of conjugated
conjugated and unconjugated linoleic acid during heating. Food Chemistry, 140(4),
linolenic acids. Journal of Agricultural and Food Chemistry, 57, 4212–4217.
680–685.
Yang, X., Gao, N., Hu, L., Li, J., & Sun, Y. (2015). Development and evaluation of novel
Gouin, S. (2004). Microencapsulation: Industrial appraisal of existing technologies and
microcapsules containing poppy-seed oil using complex coacervation. Journal of Food
trends. Trends in Food Science & Technology, 15, 330–347.
Engineering, 161, 87–93.
Goula, A. M., & Adamopoulos, K. G. (2012). A method for pomegranate seed application
in food industries: Seed oil encapsulation. Food and Bioproducts Processing, 90,

10