International Journal of Biological Macromolecules 79 (2015) 894–902

Contents lists available at ScienceDirect

International Journal of Biological Macromolecules

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

Characterization of fish gelatin–gum arabic complex coacervates

as influenced by phase separation temperature
Mohammad Anvari a,b , Cheol-Ho Pan c,d , Won-Byong Yoon e , Donghwa Chung f,g,∗
a
Department of Marine Food Science and Technology, Gangneung-Wonju National University, Gangneung 210-702, Republic of Korea
b
School of Food Science, University of Idaho, Moscow, ID 83843, USA
c
Laboratory of Biomodulation, KIST Gangneung Institute of Natural Products, Gangneung 210-340, Republic of Korea
d
Department of Biological Chemistry, Korea University of Science and Technology (UST), Daejeon 305-350, Republic of Korea
e
Department of Food Science and Biotechnology, Kangwon National University, Chuncheon 200-701, Republic of Korea
f
Graduate School of International Agricultural Technology, Seoul National University, Pyeongchang 232-916, Republic of Korea
g
Institute of Food Industrialization, Institutes of Green Bio Science and Technology, Seoul National University, Pyeongchang 232-916, Republic of Korea

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

Article history: The rheological and structural characteristics of fish gelatin (FG)–gum arabic (GA) complex coacervate
Received 29 March 2015 phase, separated from an aqueous mixture of 1% FG and 1% GA at pH 3.5, were investigated as influenced
Received in revised form 30 May 2015 by phase separation temperature. Decreasing the phase separation temperature from 40 to 10 ◦ C lead to:
Accepted 2 June 2015
(1) the formation of a coacervate phase with a larger volume fraction and higher biopolymer concentra-
Available online 6 June 2015
tions, which is more viscous, more structural resistant at low shear rates, more shear-thinning at high
shear rates, and more condensed in microstructure, (2) a solid-like elastic behavior of the phase separated
Keywords:
at 10 ◦ C at a high oscillatory frequency, (3) the increase in gelling and melting temperatures of the coacer-
Fish gelatin–gum arabic complex
coacervation
vate phase (3.7−3.9 ◦ C and 6.2−6.9 ◦ C, respectively), (4) the formation of a more rigid and thermo-stable
Phase separation temperature coacervate gel. The coacervate phase is regarded as a homogeneously networked biopolymer matrix dis-
Oscillatory shear persed with water vacuoles and its gel as a weak physical gel reinforced by FG–GA attractive electrostatic
interactions.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction two-phase system where one phase is enriched in insoluble

Protein–polysaccharide interactions in an aqueous environment
have been extensively studied in recent years because of their
high potential applications in the development of bioactive deliv-
ery devices, fat replacers, meat analogs, gels, emulsions, edible
films, and coatings in food and pharmaceutical industries [1–3].
The interactions are either attractive or repulsive, depending on
biopolymer characteristics (e.g., molecular weight, charge den-
sity, conformation, and flexibility), solvent properties (e.g., pH and
ionic strength), and mixing conditions (e.g., protein to polysaccha-
ride ratio, total biopolymer concentration, temperature, stirring,
and pressure) [4,5]. Attractive interactions, primarily attributed
to electrostatic attractions between oppositely charged biopoly-
mers, induce the formation of biopolymer complexes, which
could be either soluble in a single phase or insoluble to build a
∗ Corresponding author at: Graduate School of International Agricultural Tech-
nology, Seoul National University, Pyeongchang 232-916, Republic of Korea.
Tel.: +82 33 339 5793; fax: +82 33 339 5716.
E-mail address: dchung@snu.ac.kr (D. Chung).
complexes and the other depleted [3,6]. The sequential process,
composed of (1) the formation of charge neutralized insoluble com-
plexes and (2) the subsequent spontaneous macroscopic phase
separation to complex-rich and solvent-rich phases, is known as
associative separation, which can be also called complex coacer-
vation or precipitation for the separation of liquid-state insoluble
complexes (complex coacervates) or solid-state insoluble com-
plexes (precipitates), respectively [3,6].
The formation of insoluble protein–polysaccharide complexes
can be significantly influenced by temperature through the
temperature-dependent alterations not only in the conformation
of biopolymers but also in the noncoulombic interactions between
the biopolymers, such as hydrogen bonding and hydrophobic inter-
actions. Harding et al. [7] demonstrated that the conformational
change of bovine serum albumin by heat-induced denaturation
enhanced the formation of insoluble complexes with alginate. Fang
et al. [8] showed that the conformational change of pigskin gelatin
by heat-induced transition from helix to random coil could also
enhance the complexation with ␬-carrageenan. Hydrogen bonding,
in principle, is favored at low temperatures, whereas hydropho-
bic interactions become stronger with increasing temperature,

http://dx.doi.org/10.1016/j.ijbiomac.2015.06.004
0141-8130/© 2015 Elsevier B.V. All rights reserved.
 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902
because of the exposure of more interaction sites by heat-induced
conformational changes of the biopolymers [1,9,10]. Liu et al. [11]
also reported that a decrease of temperature from 23 to 6 ◦ C sig-
nificantly improved the complexation between pea protein isolate
and gum arabic by strengthening hydrogen bonding as a sec-
ondary force, whereas hydrophobic interactions appeared to be
more prevalent in a higher temperature range of 23−60 ◦ C and
to be associated with the stability, rather than the formation of
complexes.
The spontaneous macroscopic phase separation of insoluble
complexes may be also strongly dependent on temperature. The
phase separation phenomenon, mainly driven by entropy, takes
place by a nucleation-and-growth-like mechanism, as a result of the
rearrangement, coalescence, Ostwald ripening, or aggregation of
insoluble complexes, mostly via the reformation of intermolecular
interactions and the expulsion of water molecules [1,12,13]. It was
suggested by Nigen et al. [14] that the initial complexation between
two oppositely charged globular proteins, ␣-lactalbumin and
lysozyme, might be caused by electrostatic attractions, however,
the growth of complexes could be largely driven by hydrogen bond-
ing at low temperatures, resulting in the formation of reversible
aggregates, but by hydrophobic interactions at higher temperatures
to form coacervate-like structures. Kayitmazer et al. [15] showed
that the shear-dependent viscosity and dynamic moduli of bovine
serum albumin-chitosan coacervate phase was significantly depen-
dant on temperature, although the temperature was not the phase
separation temperature but the temperature of measurements. It is
therefore expected that the rheology and structure of complex-rich
phase obtained by protein–polysaccharide complex coacervation
and the sol–gel transformation of such coacervate phase, which are
directly related to the success of some valuable applications such as
microencapsulation and emulsion stabilization, could be consider-
ably influenced by the temperature of phase separation. However,
such aspects of complex coacervation have not been systematically
studied.
Fish gelatin (FG) is regarded as a promising alternative to mam-
malian gelatins, because it can be produced from fish skin and
bones, comprising nearly 30% of fish processing byproducts, and
does not have safety- or religion-related consumer concerns [3].
It was found by our group that the FG from cold water fish skin
(58 kDa) underwent complex coacervation with gum arabic (GA,
382 kDa), one of the most widely used anionic polysaccharides,
in aqueous solutions at 40 ◦ C [3]. The formation of insoluble or
soluble FG–GA complexes was strongly influenced by pH and FG
to GA weight ratio, as often reported for other protein–anionic
polysaccharide pairs, and the total concentration and composition
of biopolymers in the coacervate phase was dependent on the total
concentration of biopolymers in the initial mixture. In the present
study, we aimed to investigate the rheological and structural
characteristics of FG–GA complex coacervate phase as influenced
by phase separation temperature, using dynamic oscillatory and
steady shear rheological measurements, Fourier transform infrared
spectroscopy (FTIR), confocal scanning laser microscopy (CSLM),
and scanning electron microscopy (SEM).
2. Materials and methods
2.1. Materials
Fish gelatin (FG, from cold water fish skin) and gum arabic (GA,
from Acacia tree) were purchased from Sigma Chemical Co. (St.
Louis, MO, USA) and Carl Roth GmbH (Karlsruhe, Germany), respec-
tively. The average molecular weights of FG and GA, determined
by viscometry using an Ostwald viscometer and Mark–Houwink
equation in our preliminary experiments, were 58 and 382 kDa,
895
respectively [3]. Sodium azide (NaN3 ) was obtained from Daehung
Chemicals & Metals (Siheung, Korea). Fluorescein isothiocyanate
(FITC), rhodamine B isothiocyanate (RITC), and dimethyl sulfoxide
(DMSO) were purchased from Sigma Chemical Co. (St. Louis, MO,
USA). All other reagents were of analytical grade purity.
2.2. Preparation of FG–GA coacervate phase
Accurately weighed amounts of FG and GA were separately dis-
solved in 100 mL of distilled water at 40 ◦ C and mixed together at a
weight ratio of FG to GA (FG:GA) of 1:1 and a total biopolymer con-
centration of 2% (w/v), followed by adjusting the pH to 3.5. Sodium
azide was added as a preservative at a concentration of 0.02% (w/v).
The FG–GA mixture was incubated at 40 ◦ C in a shaking water bath
at 100 rpm for 24 h for sufficient electrostatic interactions between
FG and GA molecules, and then placed statically at 10, 30, or 40 ◦ C
for another 24 h for the macroscopic phase separation of FG–GA
insoluble complexes to reach equilibrium state.
2.3. Measurements of phase volume and composition
The volume fraction of coacervate phase was determined as
follows:
Vc
Volume fraction (%) = × 100 (1)
Vt
where Vc is the volume of coacervate phase and Vt is the total vol-
ume of FG–GA mixture. The amounts of total biopolymers and FG
in solvent-rich phase were measured by oven drying at 105 ◦ C for
24 h and by the Lowry method, respectively [3,16]. The values mea-
sured above were used to calculate the amount and concentration
of biopolymers in the coacervate phase and the FG:GA value of
coacervate phase. All measurements were performed at least in
triplicate.
2.4. Fourier transformed infrared (FTIR) spectroscopy
FTIR spectroscopy was used to examine the changes in the
secondary structure of FG and molecular interactions between
biopolymers, such as electrostatic interactions, hydrogen bonding,
or hydrophobic interactions, in the coacervate phase. The coacer-
vate phases obtained were dried in a vacuum oven at their phase
separation temperatures. Aqueous solutions of FG and GA, sepa-
rately prepared at 1% (w/v) and pH 3.5, were also dried as controls.
The dried samples of 2 mg in approximately 100 mg potassium
bromide were placed in discs, and FTIR spectra were obtained in
the wavenumber range of 4000–500 cm−1 using an infrared spec-
trophotometer (Tensor 27, Bruker Instruments, Billerica, MA, USA)
at a scanning rate of 10 kHz with 64 scans and a resolution of 4 cm−1 .
The background was obtained against pure potassium bromide pel-
let. All data treatments were carried out using OPUS 7.0 software
(Bruker Instruments, Billerica, MA, USA).
2.5. Confocal scanning laser microscopy (CSLM)
CSLM was used to examine the microstructure of FG–GA coac-
ervate phase. FG and GA were covalently stained with fluorescent
markers, FITC and RITC, respectively. The excitation/emission
wavelengths of FITC and RITC are 485/530 nm and 530/590 nm,
respectively. The staining was performed as described in Schmitt
et al. [9]. Briefly, aqueous solutions of FG and GA were separately
prepared at 2% (w/v), and their pH values were adjusted to 8.5
and 10.5, respectively, to facilitate the staining reaction. There-
after, 25 ␮L of 2% (w/w) FITC or RITC solution in DMSO was added
to 100 mL of the FG or GA solution, respectively, and the staining
reaction was allowed to proceed for 1.5 h with gentle stirring at
896 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902

Table 1 Secondly, the dynamic oscillatory behavior of the three FG–GA

Volume fraction and composition of FG–GA coacervate phasesa separated at three
coacervate phases was examined by frequency sweep tests, where
different temperatures.
the elastic or storage modulus (G ) and viscous or loss modulus (G )
Parameter Phase separation temperature of each phase were determined as a function of angular frequency
40 ◦ C 30 ◦ C 10 ◦ C (ω) in the range of 0.1–100 rad/s. The linear viscoelastic region,
where G and G are independent of strain (), was determined by
Volume fraction (%, v/v) 7.0 ± 0.1b
7.1 ± 0.3b
7.8 ± 0.1a
FG content (mg) 233.3 ± 1.8b 235.2 ± 1.3b 239.0 ± 0.8a strain sweep tests in the  range of 0.01–100% and at ω = 2 rad/s.
GA content (mg) 170.5 ± 1.3c 174.9 ± 0.8b 231.0 ± 2.0a For the coacervate phases separated at 10, 30, and 40 ◦ C, the lin-
Total biopolymer content (mg) 403.8 ± 1.9c 410.0 ± 2.5b 470.0 ± 1.7a ear viscoelastic region was found to be up to 2.0%, 1.3%, and 1.2%,
FG concentration (%, w/v) 8.3 ± 0.0a 8.3 ± 0.0a 7.7 ± 0.0b respectively (data not shown), and therefore, a  value of 1% was
GA concentration (%, w/v) 6.1 ± 0.0c 6.2 ± 0.0b 7.4 ± 0.0a
Total biopolymer 14.4 ± 0.1b 14.4 ± 0.1b 15.1 ± 0.1a
used for the frequency sweep and temperature sweep tests.
concentration (%, w/v) Thirdly, the sol–gel transformation of the three FG–GA coacer-
FG to GA weight ratio (FG:GA) 1.4 ± 0.0a 1.3 ± 0.0a 1.0 ± 0.1b vate phases was investigated by temperature sweep tests, where G
a
All the phases were separated from the FG–GA mixture prepared at a total and G were monitored at ω = 2 rad/s and  = 1% with decreasing
biopolymer concentration of 2% (w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C. Values are temperature from each phase separation temperature to 3 ◦ C and
expressed as mean ± standard deviation of at least three different samples. Val- then back to the phase separation temperature at a cooling or heat-
ues with different superscript letters in the same row are significantly different at ing rate of 0.016 ◦ C/min, and the gelling and melting points (Tg and
p < 0.05.
Tm , respectively) of each coacervate phase were determined. Com-
plex modulus (G*) was obtained at 3 ◦ C and used to compare the
overall rigidity of the gels transformed from the coacervate phases.

40 ◦ C. The two solutions were then mixed together as described 

in Section 2.2 to obtain FG–GA coacervate phases separated at
three different temperatures (10, 30, and 40 ◦ C). The microstruc-
ture of stained FG–GA coacervate phases were imaged using FV-300
confocal scanning unit (Olympus, Tokyo, Japan) equipped with an
Olympus IX71 inverted microscope and an argon ion laser. The
laser was adjusted in the green/red fluorescence mode, yielding
two excitation wavelengths of 485 and 530 nm, to obtain green and
red fluorescence images from two separate channels. All images
were taken at a magnification of 40× (oil immersion, numeric aper-
ture 1.30) and processed using Fluoview Software (Olympus, Tokyo,
Japan).
G∗ = (G )2 + (G )2 (2)
Finally, the dynamic oscillatory behavior of the gels obtained by
cooling the coacervate phases to 3 ◦ C was also examined by fre-
quency sweep tests at  = 1% in the range of 0.1–100 rad/s. All the
rheological measurements were performed at least in triplicate.
3. Results and discussion
3.1. Volume fraction and composition of FG–GA coacervate phase

The volume fraction of FG–GA coacervate phase was found to

2.6. Scanning electron microscopy (SEM)
SEM was also used to examine the microstructure of the FG–GA
coacervate phases separated at three different temperatures (10,
30, and 40 ◦ C). Each sample was fixed in 2% (w/w) glutaraldehyde
solution, dehydrated using different aqueous ethanol solutions, and
placed in acetone as described in Espinosa-Andrews et al. [17]. Each
sample was dried in a vacuum oven for 5 min, fragmented, and
mounted on stubs with the fractured face upwards, and gold-coated
with a fine coat ion sputter (JFC 1100, JEOL Ltd., Akishima, Japan).
A low vacuum scanning electron microscope (Inspect F50, FEI Co.,
Hillsboro, OR, USA), was used at 15 kV to obtain the SEM images
of the cross-section of each sample at magnifications of 100× and
1000×, respectively.
2.7. Rheological analysis
All rheological measurements were carried out using a TA-
AR 2000 rheometer (TA Instruments Inc., New Castle, DE, USA),
equipped with a cone–plate geometry (diameter 40 mm, angle 2◦ ,
and gap 53 ␮m) and a moisture trap. Firstly, the steady shear behav-
ior of the FG–GA coacervate phases separated at three different
temperatures (10, 30, and 40 ◦ C) was examined by measuring the
apparent viscosity (a ) of each phase as a function of shear rate
().
˙ Each sample was sheared from 0.01 to 100 1/s at its phase sep-
aration temperature, during which 61 data points were recorded
at 6 s intervals. As controls, three aqueous mixtures of FG and
GA were prepared with the same biopolymer compositions as the
coacervate phases, determined in Section 2.3 (Table 1), at pH 8.0,
where there are no significant attractive interactions between FG
and GA.
increase from 7.0% to 7.8% with decreasing phase separation tem-
perature from 40 to 10 ◦ C (Table 1). The values were close to the
value (about 8.0%) reported for the coacervate phase separated
from the mixture of gelatin (unknown source) and GA, which was
prepared under similar conditions as in the present study: a weight
ratio of gelatin to GA of 1:1, a total biopolymer concentration of 2%,
pH about 3.5, and a mixing and phase separation temperature of
40 ◦ C [18]. The composition analysis shows that the amount and
concentration of total biopolymer and GA in the coacervate phase,
as well as the amount of FG in the phase significantly increased
with decreasing phase separation temperature, while the concen-
tration of FG was reduced from 8.3% to 7.7% (w/v), resulting in
the considerable decrease in the weight ratio of FG to GA (FG:GA)
from 1.4 to 1.0 (Table 1). The results indicate that decreasing phase
separation temperature from 40 to 10 ◦ C not only increase the vol-
ume fraction of coacervate phase, but also enhance the partition
of the two biopolymers, especially GA molecules, into the coac-
ervate phase. The volume increase of coacervate phase may be
partly attributed to the partition of more GA molecules (382 kDa),
which are much larger than FG molecules (58 kDa), into the phase at
lower phase separation temperature [13]. The phase separation of
protein–polysaccharide mixture is known to be an entropy-driven
process occurring through a nucleation-and-growth like mecha-
nism, where the biopolymer complexes formed during the mixing
undergo rearrangement, coalescence, Ostwald ripening, or aggre-
gation, mostly via the reformation of intermolecular interactions
and the expulsion of water molecules [1,12,13]. Different phase
volumes and compositions of FG–GA coacervate phases separated
at different temperatures suggest that temperature could be an
important factor influencing the phase separation mechanism of
FG–GA complexes.
 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902 897

3.2. Fourier transformed infrared (FTIR) spectroscopy 10

o
FG-GA at 40 C
o
Fig. 1 shows the FT-IR spectra of FG, GA, and the FG–GA FG-GA at 30 C
o
coacervate phases separated at three different temperatures. The FG-GA at 10 C
changes in the secondary structure of FG and the molecular interac- Control at 40 oC
1 Control at 30 oC
tions between biopolymers, such as electrostatic interactions and

ηa (Pa.s)
Control at 10 oC
hydrogen bonding, in the coacervate phases were investigated by Carreau fitting curve
examining three band regions: amide A (3600–2300 cm–1 ) for N H
stretching, amide I (1600–1700 cm–1 ) for primarily C O stretching,
and amide II (1335–1560 cm–1 ) for N H bending and C N stretch- 0.1
ing [18–21]. At 40 ◦ C, FG molecules showed a typical amide I peak
at 1635 cm−1 , which is regarded as a characteristic for the random
coil structure of gelatin [19,22,23], and GA molecules showed two
peaks at 1612 and 1423 cm−1 due to the asymmetric and symmet- 0.01
ric stretching of carboxyl acid salt COO− , respectively (Fig. 1) [17]. 0.01 0.1 1 10 100
The FG–GA coacervate phase obtained at 40 ◦ C did not show the two .
GA peaks, implying the involvement of COO− of GA molecules in γ (1/s)
electrostatic interactions, and its amide I peak at 1632 cm−1 indi- Fig. 2. Apparent viscosity (a ) versus shear rate () ˙ for FG–GA coacervate phases
cates that the FG molecules complexed with GA molecules also had separated at 10, 30, or 40 ◦ C and their corresponding controls. The coacervate phases
a random coil structure at 40 ◦ C. were separated from the FG–GA mixture prepared at a total biopolymer concentra-
At 30 ◦ C, FG showed a similar spectrum to that at 40 ◦ C, while tion of 2% (w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C. The controls are the FG–GA mixtures
prepared with the same biopolymer compositions as the coacervate phases, but at
the FG–GA coacervate phase exhibited an amide A peak at a
pH 8.0.
lower wavenumber (3420 cm−1 ) and an amide II peak at a higher
wavenumber (1539 cm−1 ) (Fig. 1). Amide A and amide II bands are
attributed to the stretching vibration of N H group bonded to O H downshift of amide A peak and the upshift of amide II peak suggest
group and the C N stretching combined with N H deformation the increase of hydrogen bonding in the coacervate phase [20,24].
of peptide group, respective, and therefore, both the wavenumber The FG molecules in the coacervate phase may also have a random
coil structure at 30 ◦ C, considering the amide I peak at 1639 cm−1 .
When the temperature was further reduced to 10 ◦ C, FG showed
amide I and amide II peaks at higher wavenumbers (1656 and
amide A

amide II

1541 cm−1 , respectively) (Fig. 1). The amide I peak at 1656 cm−1
amide I

is regarded as a characteristic for the helical structure of collagen

[23], and it is known that cold water fish gelatin undergoes a struc-
tural transition from a random coil to a helical conformation at
FG 40 oC
around 15 ◦ C [25]. The spectral changes, therefore, may indicate
3429 1635 1521 that FG molecules had an ordered and entangled helical secondary
structure at 10 ◦ C. The FG–GA coacervate phase at 10 ◦ C exhibited
FG 30 oC an amide A peak at a lower wavenumber (3410 cm−1 ) compared
to those prepared at higher temperatures (Fig. 1), suggesting that
1521
3427 1633 FG–GA coacervate phase had increased hydrogen bonding [24]
o
FG 10 C at a lower phase separation temperature. The stronger hydrogen
bonding in FG–GA coacervate phase at a lower phase separation
temperature, especially at 10 ◦ C, could be a reason for the enhanced
Transmittance (%)

partition of the biopolymers, especially GA molecules, into the

1541
1656 coacervate phase with a lower FG:GA value at a lower tempera-
o
3423 GA 40 C ture (Table 1). The FG–GA coacervate phase at 10 ◦ C also showed an
amide I peak at 1654 cm−1 (Fig. 1), indicating that the FG molecules
1612 1423 in the coacervate phase at 10 ◦ C also had a helical secondary struc-
3440
ture.
FG-GA 40 oC
3.3. Steady shear behavior of FG–GA coacervate phase
1523
1632
3427
Fig. 2 shows the apparent viscosity (a ) of the FG–GA coacer-
FG-GA 30 oC
1539 vate phases separated at three different temperatures as a function
1639
3420 of shear rate ().
˙ Three corresponding control samples, which are
the FG–GA mixtures prepared with the same biopolymer compo-
FG-GA 10 oC sitions as the coacervate phases, but at pH 8.0, where there is no
1541 significant attractive electrostatic interactions between FG and GA
1654
3410 molecules, were also investigated. At all the tested temperatures,
the coacervate phases showed much higher values of a than the
4000 3500 3000 2500 2000 1500 1000 500 controls, although the biopolymer composition was the same. This
may be because the resistance to flow was elevated due to the
Wavenumber (cm-1) FG–GA electrostatic interactions in the coacervate phases. Wein-
breck et al. [26] also reported that the coacervate phase of whey
Fig. 1. FTIR spectra of FG, GA, and FG–GA coacervate phases separated at 10, 30, or
40 ◦ C. The coacervate phases were separated from the FG–GA mixture prepared at protein and gum arabic at pH 4.0 had much higher viscosity than
a total biopolymer concentration of 2% (w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C. the control prepared at pH 7.0 with similar composition.
898 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902
Table 2
Carreau model parameters estimated for FG–GA coacervate phasesa separated at
three different temperatures and their controls.
Sample 0 (Pa s) ∞ (Pa s) N −1
˙ c (s ) R2
FG–GA at 40 ◦ C 0.250 0.046 0.36 0.0110
FG–GA at 30 ◦ C 0.327 0.061 0.38 0.0380
FG–GA at 10 ◦ C 1.246 0.150 0.43 0.0827
Control at 40 ◦ C 0.042 0.016 0.22 0.0025
Control at 30 ◦ C 0.079 0.017 0.28 0.0028
Control at 10 ◦ C 0.198 0.019 0.30 0.0031
a
All the phases were separated from the FG–GA mixture prepared at a total
biopolymer concentration of 2% (w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C. The controls
are the FG–GA mixtures prepared at pH 8.0 with the same biopolymer compositions
as the coacervate phases, determined in Table 1.
All the samples exhibited a shear-thinning flow behavior,
and therefore, the well-known Carreau model was employed to
describe the viscosity profiles in Fig. 2 [27].
0 − ∞
a = ∞ +   2 N
˙
1+ ˙ c
where 0 is the zero-shear viscosity (Pa s), i.e., the limit viscosity at
low shear rates, ∞ is the infinite-shear viscosity (Pa s), i.e., the limit
viscosity at high shear rates, ˙ c is the critical shear rate (s−1 ) mark-
ing the onset of shear-thinning region, and N is a dimensionless
exponent related to the slope of shear-thinning region. The Car-
reau model showed a good fitting to the experimental data with
high R2 values (≥0.9973), and the model parameters estimated are
listed in Table 2.
The values of 0 and ∞ for FG–GA coacervate phase
increased by about 5-fold (0.250–1.246 Pa s) and by about 3-fold
(0.046–0.150 Pa s), respectively, with decreasing phase separa-
tion temperature from 40 to 10 ◦ C (Table 2). This is probably
because decreasing the temperature not only reduced the mobil-
ity of the biopolymers but also increased the hydrogen bonding
between the biopolymers, as observed in the FT-IR analysis. The
higher mass fraction of GA, having about 6 times larger molecular
weight (382 kDa) compared to FG (58 kDa), at a lower temperature
(Table 1) may be also attributed to the increase of 0 and ∞ with
the temperature decrease. The controls also showed about 5-fold
increase in 0 (0.042–0.198 Pa s) but only a slight increase in ∞
(0.016–0.019 Pa s) with decreasing the temperature.
The value of N for FG–GA coacervate phase increased from 0.36
to 0.43 with decreasing phase separation temperature (Table 2),
indicating that the coacervate phase was not only more viscous but
also more shear-thinning at a lower temperature. The coacervate
phase had a higher GA fraction at a lower temperature (Table 1),
and it is known that polysaccharide relaxation is mainly responsible
for the shear thinning behavior of most concentrated biopolymer
mixtures [26]. In addition, the hydrogen bonding became stronger
with the temperature decrease, even leading to the formation of
helical structure in FG at 10 ◦ C, as observed in the FT-IR analysis.
Therefore, the biopolymers in the coacervate phase might undergo
more conformational changes and rearrangements at a lower tem-
perature. The controls also showed an increase in N value from 0.22
to 0.30 with the temperature decrease. At any tested temperature,
the coacervate phase was found to have a higher N value, indicat-
ing greater shear-thinning behavior than the control, although the
biopolymer composition was the same. This is probably because the
complexes formed by attractive electrostatic interactions between
FG and GA molecules, i.e., FG–GA complex coacervates, were sub-
jected to more structural rearrangements than the biopolymers
mixed without significant attractive interactions.
The value of ˙ c for FG–GA coacervate phase significantly
increased by 7.5 times (0.0110–0.0827 s−1 ) with decreasing phase
100
G' (40 oC)
G" (40 oC)
o
10 G' (30 C)
o
G" (30 C)
0.9973 G' (10 oC)
G' or G'' (Pa)
0.9989
1 G" (10 oC)
0.9999
0.9992
0.9997
0.9998 0.1
0.01
0.001
0.1 1 10 100
ω (rad/s)
Fig. 3. Storage (G ) and loss (G ) moduli versus angular frequency (ω) for FG–GA
coacervate phases separated at 10, 30, or 40 ◦ C. The coacervate phases were sepa-
(3)
rated from the FG–GA mixture prepared at a total biopolymer concentration of 2%
(w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C.
separation temperature (Table 2), implying the increase of New-
tonian plateau region at low shear rates. This is quite interesting
because the result indicates that at a lower temperature, the coac-
ervate phase was more resistant to the structural deformation at
low shear rates, although it showed stronger shear-thinning behav-
ior once the structural deformation and rearrangements occurred
above ˙ c . This is probably ascribed to the stronger hydrogen bond-
ing and the reduced molecular mobility at a lower temperature. The
controls also exhibited an increase in ˙ c from 0.0025 to 0.0031 s−1
with the temperature decrease, but not as dramatic as the coacer-
vate phase. At any tested temperature, much higher ˙ c value, i.e.,
wider low-shear Newtonian plateau, was observed for the coac-
ervate phase than for the control. This is probably because at low
shear rates, the FG–GA attractive electrostatic interactions might
disturb the deformation of GA molecules or other molecular struc-
ture, as suggested by Weinbreck et al. [26].
3.4. Oscillatory shear behavior of FG–GA coacervate phase
Fig. 3 shows the frequency sweep of the FG–GA coacervate
phases separated at three different temperatures. For the phases
separated at 30 and 40 ◦ C, the storage modulus (G ) was found to
increase with the angular frequency (ω), but to decrease when
the frequency exceeded 50 and 25 rad/s, respectively, implying
the occurrence of structural breakdown in the phases due to the
reduced intermolecular connectivity at high frequency. However,
the G of the coacervate phase at 10 ◦ C continuously increased
with the frequency without decreasing, and the coacervate phase
obtained at a lower temperature showed a higher G at any fre-
quency. The results confirm the increase of structuration in the
coacervate phase with decreasing the temperature, probably due
to the reduced molecular mobility, the stronger hydrogen bonding,
and the higher GA fraction at a lower temperature, as previously
mentioned.
The loss modulus (G ) continuously increased with the fre-
quency at all the tested temperatures, and showed larger values
than the G values, indicating the liquid-like viscous behavior of
the coacervate phases (Fig. 3). Decreasing the temperature also
raised the G value, but the increase of G was not as significant
as that of G due to the structuration, resulting in the approach of
G to G with decreasing the temperature. The coacervate phase
at 10 ◦ C even showed a crossover of G and G at a frequency of
73 rad/s, that is, exhibited liquid-like viscous behavior below the
crossover frequency and solid-like elastic behavior above it. The
 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902 899

Fig. 4. CSLM micrographs of FG–GA coacervate phases separated at 40 ◦ C (A), 30 ◦ C (B), and 10 ◦ C (C). The images were obtained from FITC-stained FG molecules. The
coacervate phases were separated from the FG–GA mixture prepared at a total biopolymer concentration of 2% (w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C. Bar represents 50 ␮m.
(For interpretation of the references to color in text near the reference citation, the reader is referred to the web version of this article.)

FG–GA complexes in the phase at 10 ◦ C might have not enough
time to relax to a more favorable structure and dissipate energy
above the crossover frequency, and consequently, act more like a
rigid gel network, in which the energy imposed by the oscillatory
strain could be temporarily stored, leading to the solid-like elastic
behavior of the coacervate phase [28].
The coacervate phase of bovine serum albumin and chitosan also
showed a crossover frequency, below which G < G (viscous) but
above which G > G (elastic), and both dynamic moduli increased
with the decrease of temperature, although the temperature was
not the phase separation temperature but the temperature of
measurements [14]. The coacervate phase of ␤-lactoglobulin and
gum arabic also showed a transition from liquid-like to gel-like
behavior with a crossover of G and G after an aging treatment
for 48 h [29]. The coacervate phase of whey protein and gum
arabic also exhibited viscous behavior (G < G ) during the fre-
quency sweep, however, no crossover frequency was reported and
the effect of temperature was not investigated [26]. Neverthe-
less, the coacervate phases of bovine serum albumin-pectin and
␤-lactoglobulin-pectin showed gel-like elastic behavior (G > G )
during the frequency sweep [30,31]. The different oscillatory
shear behaviors of the different coacervate phases reflect charac-
teristically distinct structures of protein–polysaccharide complex
coacervates.
which were found to be smallest in the coacervate phase at 10 ◦ C,
indicating that the coacervate phase at 10 ◦ C had more condensed
microstructure than the phase at 30 or 40 ◦ C. A similar CSLM
microstructure was also reported by Schmitt et al. [29] for the
coacervate phase of ␤-lactoglobulin and gum arabic, which became
denser with the loss of water vacuoles upon aging. Fig. 5 shows the
SEM micrographs of the three FG–GA coacervate phases. All the
coacervate phases displayed a sponge-like porous microstructure
due to the entrapment of water molecules within the coacervate
structure. The water vacuoles in the coacervate microstructure, i.e.,
the pores in the micrographs, were observed to become smaller
and more homogeneous in size with decreasing the temperature,
indicating more compact coacervate microstructure at a lower tem-
perature.
According to our observation, therefore, the FG–GA coacervate
phase is regarded as a dense and coalesced liquid phase, where FG
and GA molecules are homogeneously complexed and networked
by electrostatic attractions and noncoulombic interactions, such as
hydrogen bonding, to form a continuous matrix in which water
vacuoles are dispersed. In the present study, a close relationship
was found between the microstructure and the rheological prop-
erties of the coacervate phase. The denser the microstructure is,
the more viscous, the more shear-thinning, and the more elastic
the coacervate phase is.

3.5. Microstructure of FG–GA coacervate phase

Fig. 4 shows the CSLM micrographs of the FG–GA coacervate
phases separated at three different temperatures, where the FG
molecules stained with FITC were represented in green. For all the
coacervate phases, the green-colored images were similar to the
red-colored images from RITC-stained GA molecules (not shown),
demonstrating nearly homogeneous distribution of FG and GA
molecules in the coacervate phases. The black regions in the images
represent water vacuoles inside the coacervate microstructure,
3.6. Sol–gel transformation of FG–GA coacervate phase
Fig. 6 shows the temperature dependence of G and G of the
FG–GA coacervate phases separated at three different tempera-
tures. The temperature was lowered from each phase separation
temperature to 3 ◦ C and then raised back to the phase separa-
tion temperature at a rate of 0.016 ◦ C/min. During the cooling, all
the coacervate phases showed the following notable sequential
changes in the moduli; (1) a sudden increase in G immediately

Fig. 5. SEM micrographs of FG–GA coacervate phases separated at 40 ◦ C (A), 30 ◦ C (B), and 10 ◦ C (C). The coacervate phases were separated from the FG–GA mixture prepared
at a total biopolymer concentration of 2% (w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C. Bar represents 100 ␮m.
900 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902

100 of G is probably due to the adjustment of the coacervate system

G* = 75.2 ± 1.8 Pa (A)
to the cooling conditions. As the cooling proceeds, the mobility of
Tg = 3.7 ± 0.1 oC biopolymers is reduced, and the hydrogen bonding between the
Tm = 6.2 ± 0.1 oC biopolymers becomes stronger. As discussed in previous sections,
10
these could induce the conformational transition of FG molecules
from random coli to helical structure, which was reported to occur
G' or G" (Pa)

Heating
at around 15 ◦ C for cold water fish gelatin [25]. The resulting FG
1 helices may participate in the formation of junction zones for ther-
moreversible gel networks, as known for mammalian gelatins [33].
Consequently, more and stronger inter- and intra-macromolecular
connectivities might be formed with the cooling, leading to the
0.1
G' (Cooling) Cooling
remarkable increase, first in viscous fluidity (G ), and then in elastic
G" (Cooling) solidity (G ), followed by the crossover of G and G , i.e., gelation of
G' (Heating)
the coacervate phase.
G" (Heating)
0.01 Decreasing the phase separation temperature from 40 to 10 ◦ C
0 10 20 30 40 was found to slightly increase the gelling point (Tg ) of the coacer-
Temperature ( C) o vate phase from 3.7 to 3.9 ◦ C and also increase the complex modulus
(G*) obtained at 3 ◦ C from 75.2 to 85.7 Pa (Fig. 6). This indicates
100 that the coacervate phase obtained at a lower temperature had a
G* = 79.9 ± 1.8 Pa (B) higher tendency for gelation and formed a more rigid gel, which
Tg = 3.8 ± 0.0 oC
may be also attributed to the reduced molecular mobility and the
Tm = 6.6 ± 0.0 oC stronger hydrogen bonding in the coacervate phase. In particular,
10
the coacervate phase separated at 10 ◦ C, which is below the random
coil-to-helix transition temperature of FG molecules, was found to
G' or G" (Pa)

Heating contain helical FG molecules in our FT-IR analysis, and therefore, the
1 cooling could easily associate the helices for the formation of junc-
Cooling tion zones, leading to the stronger and faster gelation, compared to
the coacervate phases separated at higher temperatures.
During the heating of the coacervate phases, the following
0.1
G' (Cooling) Heating sequential changes were observed in the moduli; (1) a crossover
G" (Cooling) of G and G between 6.2 and 6.9 ◦ C with sharp decreases in both G
G' (Heating) and G , (2) a sudden change of G from a steep to a gradual decrease
0.01
G" (Heating)
between 9.0 and 10.3 ◦ C, (3) a sudden change of G from a steep to
0 5 10 15 20 25 30 a gradual decrease between 9.8 and 10.6 ◦ C (not observed for the
coacervate phase separated at 10 ◦ C). The crossover temperature,
Temperature (oC)
at which G went below G , was defined as the melting point (Tm )
100 of the coacervate phase [32]. On the contrary to the case of cool-
G* = 85.7 ± 2.4 Pa (C) ing, the heating increases the molecular mobility and weakens the
Heating
hydrogen bonding between the biopolymers. These could loosen
Tm = 6.9 ± 0.1 oC
the junction zones of gel networks, stabilized by hydrogen bonding,
10
and shift the conformation of FG molecules from an ordered helical
to a disordered random coil state. Consequently, the heating might
G' or G" (Pa)

Tg = 3.9 ± 0.0 oC
reduce and weaken the inter- and intra-macromolecular connec-
1 Cooling
tivities, leading to the melting of the coacervate gel networks and
the dramatic losses in both elastic solidity (G ) and viscous fluid-
ity (G ). Decreasing the phase separation temperature from 40 to
10 ◦ C was found to increase the melting point (Tm ) from 6.2 to 6.9 ◦ C
0.1
G' (Cooling) (Fig. 6), also indicating that the coacervate phase separated at a
G" (Cooling) lower temperature formed a more rigid and thermo-stable gel.
G' (Heating) The Tm values were higher than the Tg values for all the three
G" (Heating)
0.01 coacervate phases due to the thermal hysteresis, which is a com-
2 4 6 8 10 monly observed indication of reluctance to the thermoreversible
o sol–gel transformation of a polymeric system. Decreasing the phase
Temperature ( C)
separation temperature from 40 to 10 ◦ C increased the difference
Fig. 6. Changes in storage (G ) and loss (G ) moduli during cooling and heating of between gelling and melting points (Tm –Tg ) from 2.5 to 3.0 ◦ C, also
FG–GA coacervate phases separated at 40 ◦ C (A), 30 ◦ C (B), and 10 ◦ C (C) at a rate indicating the formation of a more ordered and thermo-stable gel
of 0.016 ◦ C/min. The coacervate phases were separated from the FG–GA mixture by the coacervate phase separated at a lower temperature.
prepared at a total biopolymer concentration of 2% (w/w), FG:GA = 1:1, pH 3.5, and
It would be also worthwhile to examine the effect of GA, a
40 ◦ C. The moduli were monitored at ω = 2 rad/s and  = 1%, and complex modulus
(G*) was obtained at 3 ◦ C. nongelling biopolymer, on the sol–gel transformation. The temper-
ature dependence of G and G was also examined for the following
two controls: (1) an aqueous FG–GA mixture prepared at pH 8.0,
after the cooling started, (2) a rapid increase in G between 6.5 and where there is no significant FG–GA attractive interactions, with
7.7 ◦ C, (3) a second sharp increase in G between 5.2 and 5.9 ◦ C, and the same biopolymer compositions as the coacervate phase sepa-
(4) a crossover of G and G between 3.7 and 3.9 ◦ C. The crossover rated at 10 ◦ C (7.7% FG and 7.4% GA, Table 1), (2) an aqueous FG
temperature, at which G exceeded G , was defined as the gelling solution prepared at the same concentration as that in the coacer-
point (Tg ) of the coacervate phase [32]. The initial sharp increase vate phase separated at 10 ◦ C (7.7%). The FG–GA mixture prepared
 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902
Fig. 7. Storage (G ) and loss (G ) moduli (A) and tan ı (B) versus angular frequency
(ω) for the gels obtained by cooling the coacervate phases separated at 10, 30, or 40 ◦ C
to 3 ◦ C. The coacervate phases were separated from the FG–GA mixture prepared at
a total biopolymer concentration of 2% (w/w), FG:GA = 1:1, pH 3.5, and 40 ◦ C.
at pH 8.0 showed a higher gelling point (3.7 ◦ C) than the FG solution
(3.1 ◦ C), and also higher G values, indicating that the presence of
GA molecules facilitated the gelation of FG molecules. This could
be explained that the volume exclusion effect between the two
biopolymers, where the free volume of FG molecules is reduced
due to their thermodynamic incompatibility with GA molecules,
increased the probability of association of FG molecules, and the
resulting filler effect of the dispersed GA-rich phase reinforced the
FG network and increased the modulus [34].
The FG–GA mixture prepared at pH 8.0 had a gelling point close
to that (3.9 ◦ C) of the coacervate phase separated at 10 ◦ C, but much
lower G values. Its complex modulus (G*) at 3 ◦ C was measured to
be 36.7 Pa, which was only 43% of that (85.7 Pa) of the coacervate
phase at 10 ◦ C. The results clearly demonstrated that the presence of
GA molecules could significantly reinforce the gel network by their
attractive electrostatic interactions with FG molecules, much more
than by their filler effect resulting from the thermodynamic incom-
patibility between the two biopolymers. In the coacervate phase,
the GA molecules electrostatically interacted with FG molecules
seem to participate in the formation of gel network together with
FG molecules to form more compact network structure.
The coacervate phase separated at 10 ◦ C showed a higher com-
plex modulus (G*) at 3 ◦ C than the coacervate phases separated at
higher temperatures (Fig. 6), although it had a lower FG concentra-
tion (Table 1). This may be attributed not only to the existence of FG
helices at 10 ◦ C, as previously mentioned, but also to the stronger
FG–GA electrostatic interactions in the coacervate phase separated
901
at 10 ◦ C, having a FG:GA value of 1.0 (Table 1), at which the FG–GA
electrostatic interactions at pH 3.55 was reported to be the most
intense [3].
3.7. Viscoelastic behavior of FG–GA coacervate gel
Fig. 7 shows the frequency sweep (at  = 1%) of the three
FG–GA coacervate gels obtained by cooling the coacervate phases
separated at three different temperatures to 3 ◦ C. All the three coac-
ervate gels exhibited a slight increase in storage modulus (G ) and
a sharp increase in loss modulus (G ) with increasing angular fre-
quency (ω) (Fig. 7A). The values of G were higher than those of G
at any frequency, and the crossover of G and G was not observed.
Accordingly, the loss tangent (tan ı = G /G ), an indicative of liquid-
like behavior, showed a steep increase with frequency but did not
exceed unity (Fig. 7B). The results imply that the three FG–GA
coacervate gels can be classified as a weak physical gel formed by
noncovalent interactions, such as hydrogen bonding, hydrophobic
interactions, and electrostatic interactions [35], having a tendency
toward more liquid-like behavior at a higher frequency due to the
breakdown of network structure. It was also found that the gel
obtained from the coacervate phase separated at a lower temper-
ature had a smaller loss tangent at any frequency (Fig. 7B), also
demonstrating that reducing the phase separation temperature can
form a more rigid FG–GA coacervate gel network, as previously
discussed in the section of sol–gel transformation.
4. Conclusions
The present study demonstrated that decreasing the phase sep-
aration temperature from 40 to 10 ◦ C significantly influenced the
phase volume, composition, steady and oscillatory shear behaviors,
microstructure, and sol–gel transformation of FG–GA coacervate
phase, as well as the viscoelastic behavior of FG–GA coacervate
gel. The decrease in the temperature increased the volume frac-
tion of coacervate phase, and enhanced the partition of the two
biopolymers, especially GA molecules, into the coacervate phase.
This is probably because the hydrogen bonding, which is known to
be majorly responsible for the reversible aggregation of biopolymer
complexes during the nucleation-and-growth like process of phase
separation, became stronger with decreasing the phase separa-
tion temperature, as shown by FT-IR spectroscopy. The coacervate
phase became more viscous and more resistant to the structural
deformation at low shear rates with decreasing the phase sepa-
ration temperature, but showed stronger shear-thinning behavior
once the structural deformation and rearrangements occurred
above ˙ c . The decrease in phase separation temperature signif-
icantly increased the G of coacervate phase in the frequency
sweep test, and the phase separated at 10 ◦ C even showed a
solid-like elastic behavior at a frequency of above 73 rad/s. In addi-
tion, the coacervate phase was found to have more condensed
microstructure at a lower temperature. The decrease in phase sepa-
ration temperature increased the values of Tg , Tm , and (Tm –Tg ) but
decreased the tan ı of the FG–GA coacervate gels formed at 3 ◦ C,
indicating that the coacervate phase separated at a lower temper-
ature had a higher tendency for gelation and formed a more rigid
and thermo-stable gel. The major reasons for all the above results
may be that decreasing the phase separation temperature reduced
the mobility of the biopolymers, increased the mass fraction of
GA molecules, strengthened the hydrogen bonding between the
biopolymers, and formed FG helical structure at 10 ◦ C, as shown
by FT-IR spectroscopy. The FG–GA coacervate gels were classi-
fied as a weak physical gel formed by noncovalent interactions,
and the presence of GA molecules was found to significantly rein-
force the gel network probably due to their attractive electrostatic
902 M. Anvari et al. / International Journal of Biological Macromolecules 79 (2015) 894–902
interactions with FG molecules. The results obtained in the cur-
rent study provide basic knowledge necessary for the use of
FG–GA complex coacervation in many useful applications, such as
microencapsulation, textural modification, emulsion stabilization,
hydrogel formation, and meat analog development.
Acknowledgement
This research was supported by the Korea Sea Grant Program
(Gangwon Sea Grant) funded by the Ministry of Oceans and Fish-
eries in Korea.
References
[1] S.L. Turgeon, C. Schmitt, C. Sanchez, Curr. Opin. Colloid Interface Sci. 12 (2007)
166–178.
[2] F. Weinbreck, R. de Vries, P. Schrooyen, C.G. de Kruif, Biomacromolecules 4
(2003) 293–303.
[3] Y. Yang, M. Anvari, C.H. Pan, D. Chung, Food Chem. 135 (2012) 555–561.
[4] D. McClements, J. Biotechnol. Adv. 24 (2006) 621–625.
[5] C. Schmitt, S.L. Turgeon, Adv. Colloid Interface Sci. 167 (2011) 63–70.
[6] H.G. Bungenberg de Jong, in: H.R. Kruyt (Ed.), Colloid Science, vol. II, Elsevier,
Amsterdam, The Netherlands, 1949, pp. 232–258.
[7] S. Harding, K. Jumel, R. Kelly, E. Gudo, J.C. Horton, J.R. Mitchell, in: K.D.
Schwenke, R. Mothes (Eds.), Food Protein, Structure, and Functionality, VCH,
Weinheim, Germany, 1993, pp. 216–226.
[8] Y. Fang, L. Li, C. Inoue, L. Lundin, I. Appelqvist, Langmuir 22 (2006) 9532–9537.
[9] C. Schmitt, C. Sanchez, S. Desobry-Banon, J. Hardy, Crit. Rev. Food Sci. Nutr. 38
(1998) 689–753.
[10] C. Schmitt, C. Sanchez, A. Lamprecht, D. Renard, C.-M. Lehr, C.G. de Kruif, J.
Hardy, Colloids Surf. B 20 (2001) 267–280.
[11] S. Liu, Y.-L. Cao, S. Ghosh, D. Rousseau, N.H. Low, M.T. Nickerson, J. Agric. Food
Chem. 58 (2010) 552–556.
[12] C. Schmitt, L. Aberkane, C. Sanchez, in: G.O. Philips, P.A. Williams (Eds.),
Handbook of Hydrocolloids, 2nd ed., CRC Press, Boca Raton, USA, 2009, pp.
421–476.
[13] F. Weinbreck, R.H. Tromp, C.G. de Kruif, Biomacromolecules 5 (2004)
1437–1445.
[14] M. Nigen, T. Croguennec, D. Renard, S. Bouhallab, Biochemistry 46 (2007)
1248–1255.
[15] A.B. Kayitmazer, S.P. Strand, C. Tribet, W. Jaeger, P.L. Dubin,
Biomacromolecules 8 (2007) 3568–3577.
[16] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, J. Biol. Chem. 193 (1951)
265–275.
[17] H. Espinosa-Andrews, O. Sandoval-Castilla, H. Vázquez-Torres, E.J.
Vernon-Carter, C. Lobato-Calleros, Carbohydr. Polym. 79 (2010) 541–546.
[18] D.J. Burgess, J.E. Carless, J. Colloid Interface Sci. 98 (1984) 1–8.
[19] M. Ahmad, S. Benjakul, Food Hydrocolloids 25 (2011) 381–388.
[20] S. Benjakul, K. Oungbho, W. Visessanguan, Y. Thiansilakul, S. Roytrakul, Food
Chem. 116 (2009) 445–451.
[21] J.H. Muyonga, C.G.B. Cole, K.G. Duodu, Food Chem. 86 (2004) 325–332.
[22] M.H. Uriarte-Montoya, H. Santacruz-Ortega, F.J. Cinco-Moroyoqui, O.
Rouzaud-Sández, M. Plascencia-Jatomea, J.M. Ezquerra-Brauer, Food Res. Int.
44 (2011) 3243–3249.
[23] K.J. Payne, A. Veis, Bipolymers 27 (1988) 1749–1760.
[24] Z. Dong, Q. Wang, Y. Du, J. Membr. Sci. 280 (2006) 37–44.
[25] I.J. Haug, K.I. Draget, in: G.O. Philips, P.A. Williams (Eds.), Handbook of
Hydrocolloids, 2nd ed., CRC Press, Boca Raton, USA, 2009, pp. 142–163.
[26] F. Weinbreck, R.H.W. Wientjes, H. Nieuwenhuijse, G.W. Robijn, C.G. de Kruif, J.
Rheol. 48 (2004) 1215–1228.
[27] M.A. Rao, in: M.A. Rao (Ed.), Rheology of Fluid and Semisolid Foods, 2nd ed.,
Springer, New York, USA, 2007, pp. 27–58.
[28] J. Ma, Y. Lin, X. Chen, B. Zhao, J. Zhang, Food Hydrocolloids 38 (2014)
119–128.
[29] C. Schmitt, E. Kolodziejczyk, M.E. Leser, in: E. Dickinson (Ed.), Food Colloids:
Interactions, Microstructure and Processing, The Royal Society of Chemistry,
Cambridge, UK, 2005, pp. 284–300.
[30] X. Wang, J. Lee, Y.-W. Wang, Q. Huang, Biomacromolecules 8 (2007) 992–997.
[31] Q. Ru, Y. Wang, J. Lee, Y. Ding, Q. Huang, Carbohydr. Polym. 88 (2012)
838–846.
[32] M. Gudmundsson, J. Food Sci. 67 (2002) 2172–2176.
[33] C. Michon, G. Cuvelier, P. Relkin, B. Launay, Int. J. Biol. Macromol. 20 (1997)
259–264.
[34] S.R. Monteiro, C. Tavares, D.V. Evtuguin, N. Moreno, J.A.L. da Silva,
Biomacromolecules 6 (2005) 3291–3299.
[35] M.H. Tunick, J. Agric. Food Chem. 59 (2011) 1481–1486.