Rheology and microstructure of cross-linked waxy maize starch ...

Rheology and microstructure of cross-linked waxy maize starch/whey protein
suspensions
H. Vu Dang a , C. Loisel a, *, A. Desrumaux a,1 , J.L. Doublier b
a Laboratoire GEPEA, ENITIAA rue de la Géraudière, BP 82225, 44322 Nantes Cedex 3, France
b INRA, UR 1628, Unité Biopolymères, Interactions, Assemblages, 44000 Nantes, France
article info
Article history:
Received 18 July 2008
Accepted 6 January 2009
Keywords:
Cross-linked starch
Whey protein
Gel
Rheological properties
Flow properties
Viscoelastic behaviour
1. Introduction
abstract
Starch is widely used to control the structure and stability of
processed food products due to its thickening and gelling properties
after gelatinisation. When a starch-in-water dispersion is
heated above the gelatinisation temperature, starch granules loose
crystallinity, absorb a large amount of water and swell up to many
times their original size leading to the enhancement of viscosity.
Cross-linked waxy maize starch (CWMS) is a high amylopectin
starch (almost free of amylose), which has been chemically modified
to make starch granules to resist to thermal and mechanical
treatment. A gelatinized CWMS dispersion can be merely described
as a suspension of swollen particles. These systems exhibit a shearthinning
flow behaviour and a solid-like viscoelastic behaviour,
these properties being mostly governed by the volume fraction and
the deformability of swollen starch granules depending upon
* Corresponding author. Tel.: þ33 2 51 78 54 64; fax: þ33 2 51 78 54 55.
E-mail address: catherine.loisel@enitiaa-nantes.fr (C. Loisel).
1 Deceased on November 2006.
0268-005X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodhyd.2009.01.004
Food Hydrocolloids 23 (2009) 1678–1686
Contents lists available at ScienceDirect
Food Hydrocolloids
journal homepage: www.elsevier.com/locate/foodhyd
The objective of the present work was to investigate the effect of the heating process on the structural
and rheological properties of whey protein isolate/cross-linked waxy maize starch (WPI/CWMS) blends
depending upon the concentration and the starch/whey protein ratio. Starch concentration ranged from
3 to 4% (w/w) and the protein content was of 0.5, 1 and 1.5% (w/w). The blend (pH 7, 100 mM ionic
strength) was heated using a jacketed vessel at two pasting temperatures: 90 and 110 C. The particle size
distribution of the WPI suspension (1.5%) displayed three distinct classes of aggregates (0.3, 65 and
220 mm), whereas the size of swollen starch granules varied from 48 to 56 mm according to the pasting
temperature. When the two components were mixed together, the peak attributed to swollen starch
granules was attenuated and broadened towards higher values (up to 88 mm) due to protein aggregates
(260–410 mm). This effect was more pronounced as the protein concentration increased. When compared
to starch alone, the rheology of the mixed system was dramatically modified for the flow behaviour as
well as for the viscoelastic properties which changed from a solid-like (3–4% starch) to a liquid-like
behaviour (3–4% starch/1.5% protein). Microscopic observations showed aggregated proteins located in
the continuous phase and swollen starch granules as the dispersed phase. Protein aggregates were of
different sizes, part of them appeared adsorbed onto swollen starch granules while another part was
unevenly distributed in the continuous phase, yielding discontinuous network which could explain the
peculiar viscoelastic behaviour of such suspensions.
Ó 2009 Elsevier Ltd. All rights reserved.
concentration (Evans & Haisman, 1979; Evans & Lips, 1992; Loisel,
Tecante, & Doublier, 2002; Nayouf, Loisel, & Doublier, 2003; Steeneken,
1989; Tecante & Doublier, 1999). It has been shown that
within the pasting temperature range 90–110 C, the rheological
behaviour of CWMS dispersions is dominated by the volume fraction
up to 3% concentration, while above 4% concentration
deformability of swollen starch granules plays a major role (Loisel
et al., 2002; Nayouf et al., 2003).
Apart from starch, whey protein isolates (WPI) can be employed
because of their unique functional properties, i.e. gelling, thickening,
foaming and water-binding capacity. These ingredients have
a high protein content (>95%) in the globular state and consist
mostly of b-lactoglobulin (>70% of total protein) and a-lactalbumin
(15–20%). Thermal gelation of globular proteins involves an initial
denaturation/unfolding step followed by aggregation to finally
form a three-dimensional network provided the concentration is
high enough. Aggregation of unfolded protein molecules occurs
through hydrophobic and/or hydrogen interactions, and formation
of disulfide bridges (Mulvihill & Kinsella, 1987). These aggregation
and gelation phenomena are strongly dependent upon protein
concentration, pH, type and concentration of ions present in the

aqueous phase. They have been reported to take place from 62 C
for a-lactalbumin to 78 C for b-lactoglobulin in the pH and ionic
conditions of milk (Bryant & Mc Clements, 1998). For b-lactoglobulin
(200 g L 1 , pH 7), the denaturation temperature has been
found to increase from 69.5 to 76.5 C while increasing the NaCl
content from 0.5 to 2 M (Boye, Alli, & Ismail, 1996).
b-Lactoglobulin has been found to form gels for concentration
above 10 g L 1 in 100 mM NaCl and pH 7 (Durand, Gimel, & Nicolai,
2002). As for isolated b-lactoglobulin, the critical concentration for
gelation to take place in the case of WPI heated to 90 C has been
found to be 11 g L 1 (Puyol, Perez, & Horne, 2001). Increasing the
ionic strength to 60 and 100 mM NaCl resulted in a lower gelling
temperature for a WPI concentration ranging from 2.5 to 10% at pH
6.7 (Puyol et al., 2001).
In the food industry, mixed systems of proteins and polysaccharides
are often used to improve the textural properties.
Consequently, investigation of gel formation of whey protein/starch
mixed system has received considerable attention, particularly on
the basis of rheological methods (Aguilera & Baffico, 1997; Fitzsimons,
Mulvihill, & Morris, 2008; Matser & Steeneken, 1997; Okechukwu
& Rao, 1997; Ravindra, Genovese, Foegeding, & Rao, 2004).
The CWMS/WPI mixtures can be described as swollen starch
granules embedded in the continuous phase constituted by the
protein network, provided the protein concentration is high
enough. If no interaction occurs between these two biopolymers,
the proteins are located in the intergranular phase between the
starch granules. Upon heating, the whey protein concentration in
the intergranular phase would increase as a result of the exclusion
by swollen starch granules and the protein contributes to the
properties of the continuous phase (Matser & Steeneken, 1997), as it
has been suggested in the case of starch/polysaccharide mixtures
(Alloncle, Lefebvre, Llamas, & Doublier, 1989; Tecante & Doublier,
1999). It is then possible to calculate the effective protein concentration
in the intergranular phase, from the volume fraction of
starch in the mixture.
The aim of this study was to describe the rheological properties
and the structure of a CWMS/WPI mixture at fixed pH (7) and ionic
strength (100 mM) and for different CWMS/WPI ratios, for two
pasting temperatures 90 and 110 C. Two starch concentrations, 3
and 4%, were used. The WPI nominal concentrations (0.5, 1.0, 1.5%)
were chosen in order to lie under (0.5%) or at the vicinity (1.5%) of
the critical gelling concentration.
2. Materials and methods
2.1. Materials
The cross-linked waxy maize starch (CWMS, adipate crosslinked
and acetate stabilized) was supplied by Roquette Frères
(Lestrem, France); the whey protein isolate WPI (Prolacta 90, 92.4%
protein) was obtained from Lactalis (Laval, France).
2.2. Preparation of the mixtures
The WPI powder was first dispersed in demineralised water at
40 C under strong mechanical stirring (1300 rpm), to obtain 0.5,
1.0 or 1.5% concentration, then submitted to milder stirring
conditions (300 rpm) for 1 h; the dispersion was then cooled down
to 25 C. NaCl was added up to 100 mM and the pH adjusted to pH 7
with 1 M NaOH; the preparation was left overnight at 4–5 C under
magnetic stirring.
Starch at 3 or 4% (w/w) was slurried in the WPI solution previously
prepared at 0.5, 1.0, or 1.5% (w/w) at 60 C under mechanical
stirring. The thermal treatment was then applied to these
suspensions in a jacketed vessel (capacity 2 L, stirring rate 400 rpm)
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686 1679
pre-heated at 90 or 110 C. This temperature was maintained for
10 min. The suspension was then cooled down to 70 C before
sampling. It is to be emphasized that the dispersion was continuously
stirred during the preparation procedure. Concentrations
were then controlled by weighing aliquots overnight at 102 C
(accuracy of the order of 0.5%).
2.3. Particle size determination
Particle size determination was performed at room temperature
using a Malvern Master sizer (Malvern Instrument, Ltd) with
a 300 mm Fourier cell (range 0.05–879 mm) using 1.529 for the
refractive index and 0.1 absorption for starch and 1.545 and 0.001
for WPI respectively. The starch paste was diluted (1/10) with
demineralised water, then dispersed in the sample dispersion unit
(1/100 ml water) and fed into the measuring cell. From each characterization,
the median volume diameter D (v, 0.5) defined as the
diameter for which 50% of the particles (by volume) are larger than
this value was chosen to allow comparisons with the literature
(Okechukwu & Rao, 1995). When the size distribution was polydisperse
as for WPI solutions or for the WPI/starch mixtures the
peaks of the particle size distribution were favoured to describe the
curves. The refractive index used for the mixture was the same as
for starch alone. Each experiment was repeated twice with an
accuracy of about 0.5%.
2.4. Rheological measurements
Flow and viscoelastic properties of the suspensions were
measured at 60 C using a controlled stress rheometer (AR1000, TA
Instruments) with the cone/plate geometry (6 cm/2 ). The exposed
edge of the sample was coated with a thin layer of paraffin oil in
order to minimize moisture loss during the measurements. For flow
curves, an up–down shear scan from 0.01 to 300 s 1 (30 s) was
applied, followed by a logarithmic stepwise decrease from 300 to
0.01 s 1 after equilibrium for each shear rate. For viscoelastic
measurements in oscillatory shear, the samples were subjected to
a frequency scan from 40 to 0.2 rad/s at 4% strain deformation
(within the linearity limits of viscoelasticity). Each experiment was
repeated twice with an accuracy of about 5%.
2.5. Microscopic observation
2.5.1. Light microscopy
Optical images were obtained on a microscope equipped with
a CCD camera (Leica Microscope). The samples were diluted and
dyed with ‘‘Fast green FCF’’ before being observed with a magnification
of 400.
2.5.2. Confocal laser scanning microscopy (CLSM)
CLSM was used in the fluorescent mode. Observations were
made with a Carl Zeiss LSM 410 Axiovert (Le Pecq, France) using
a water-immersed 40 objective. The proteins were labelled by
adding 100 mL ofa2gL 1 ANS (8-anilino-1-naphtalene sulphonic
acid) solution to 2 mL of the mixture. The labelled solution was
then treated at 95 C during 10 min, poured warm onto the slide,
and then left during 24 h at ambient temperature. ANS is known
to become fluorescent in the UV light when it is adsorbed onto
hydrophobic zones. This allows one to observe proteins without
covalently bound fluorescent labelling (Bourriot, Garnier, &
Doublier, 1999). The excitation using the UV laser was performed
at 364 nm and the emission was recorded between 450 and
497 nm.

1680
2.5.3. Phase-contrast light microscopy
The samples were observed without any dilution on a Nikon
Eclipse E 400 microscope with a 20 objective.
3. Results
3.1. Particle size distributions
Typical results for particle size distribution are shown in Fig.1–3.
Fig. 1 shows the particle size distribution curves of the WPI solution
(1.5%), the CWMS suspension (3%) and WPI/CWMS mixtures all
heated to 90 C. For WPI alone, three populations were observed
corresponding to three sizes of whey protein aggregates. The most
important fraction was centred at around 0.3 mm. It is to be noticed
that the size of the b-lactoglobulin dimer is about 6 nm. Therefore,
the size of 0.3 mm is to be ascribed to protein aggregates likely owing
to the association of primary aggregates as described by Mahmoudi
(2007) in the same pH and ionic environment. The two other peaks
were centred at w65 mm and 220 mm. They may be due to the
growth or aggregation of the former aggregates into larger ones.
For the 3% CWMS dispersion, there was only one peak (D (v,
0.5) ¼ 48 mm), representing the swollen starch granules in agreement
with previous results (D (v, 0.5), w45 mm) (Nayouf et al.,
2003). The profile displayed by the 0.5% WPI/3% CWMS mixture
was almost superimposed to that of CWMS alone. There was no
evidence of whey protein aggregates below 1 mm and the size of the
swollen starch granules appeared unchanged. There was however
a slight shift of the maximum of the peak from 48 to 56 mm. For the
1.5% WPI/3% CWMS mixture, dramatic changes were experienced:
the main peak was still centred at 56 mm but this was broadened
while a shoulder appeared at around 260 mm. This could be
ascribed to the presence of large protein aggregates in the same size
range as for 1.5% WPI solutions alone (3rd peak). It is to be noticed
that the main peak (56 mm) can be ascribed to swollen starch
granules as well as to protein aggregates (2nd peak centred at
65 mm).
Fig. 2 shows the distribution profiles with 4% CWMS at the same
pasting temperature (90 C). The pattern exhibited by the 4%
CWMS alone suspension was similar to that of the 3% CWMS
suspension with a single peak at 48 mm. For the 0.5% WPI/4% CWMS
mixture, a single peak was also present but this was shifted to
w65 mm as previously seen for the 1.5% WPI solution alone. For the
Volume Frequency (%)
20
18
16
14
12
10
8
6
4
2
0
10 -1
10 -1
1 10 10 2
Particle diameter (µm)
Fig. 1. Particle size distribution of 1.5% WPI alone ( ), 3% CWMS alone (>), 0.5%
WPI/3% CWMS (,) and 1.5% WPI/3% CWMS (D) mixtures pasted at 90 C.
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686
10 3
Volume Frequency (%)
20
18
16
14
12
10
8
6
4
2
0
10 -1
10 -1
1 10 10 2
Particle diameter (µm)
Fig. 2. Particle size distribution of 1.5% WPI alone ( ), 4% CWMS alone (>), 0.5%
WPI/4% CWMS (,) and 1.5% WPI/4% CWMS (D) mixtures pasted at 90 C.
1.5% WPI/4% CWMS mixture, the profile displayed some similarity
to that of the 1.5% WPI/3% CWMS mixture (Fig. 1) unless the peak
was centred at 88 mm and the shoulder was much more accentuated
and shifted to around 300 mm.
Fig. 3 shows the particle size distribution of the 1.5 WPI/3%
CWMS dispersion heated to 110 C. This is compared to the 3%
CWMS and the 1.5% WPI dispersions. The peak of the 3% CWMS
alone dispersion was slightly shifted from 48 to 56 mm as expected
(Nayouf et al., 2003). For 1.5% WPI alone, a unique peak at 220 mm
was observed showing that increasing the heating temperature
from 90 to 110 C dramatically modified the size distribution of
whey protein aggregates. The smallest aggregates (from 0.1 to
1 mm) seen at 90 C completely disappeared and a unique population
of about 200 mm was observed. This is in agreement with
the fact that increasing the temperature enhances the denaturation
rate of the WPI and consequently the size of the aggregates (Kazmierski
& Corredig, 2003). For the 1.5% WPI/3% CWMS mixture, two
distinct peaks centred, respectively, at 56 and 410 mm were clearly
identified instead of the broad one observed at 90 C. Table 1
Volume Frequency (%)
20
18
16
14
12
10
8
6
4
2
0
10 -2
10 -1
1 10 10 2
Particle diameter (µm)
Fig. 3. Particle size distribution of 1.5% WPI alone ( ), 3% CWMS alone (A) and 1.5%
WPI/3% CWMS (:) mixtures pasted at 110 C.
10 3
10 3

summarises the overall data in terms of peak diameters obtained in
1.5% WPI alone, 3 or 4% CWMS and mixed WPI/CWMS dispersions
heated to 90 C or 110 C. We attempted to deconvoluate the
distribution curves; unfortunately this could not be performed
accurately since the populations for Figs. 1 and 2 were too close to
each other. Finally, increasing starch concentration in the WPI/
CWMS mixture from 3 to 4% also shifted the peaks to larger values
suggesting whey protein aggregates of a larger size. From these
data, it appears that the size of the swollen starch granules was
mostly determined by the pasting temperature and apparently was
not strongly influenced by the presence of proteins.
3.2. Rheological properties
All rheological measurements have been performed at the end
of the cooling period at 60 C. All flow curves at 60 C in linear
scales of the CWMS alone suspensions as well as of the WPI/CWMS
mixtures exhibited a non-Newtonian shear-thinning behaviour
with an anti-clockwise loop (‘‘anti-thixotropic’’ behaviour) (not
showed). Such a behaviour has been reported for CWMS suspensions
(Genovese & Rao, 2003; Nayouf et al., 2003; Tecante & Doublier,
1999). This may be ascribed to a rearrangement of closepacked
deformable particles under shearing to yield a higher
organization of the system (Barnes, Hutton, & Walters, 1989).
Therefore, adding WPI to the starch suspension did not change the
anti-thixotropic character of the CWMS alone suspensions.
Figs. 4, 5 and 6 show the equilibrium flow curves obtained by
stepwise measurements in the presence of 3% starch pasted at
90 C, 4% at 90 C and 3% at 110 C, respectively. Regarding the
starch suspensions alone, the entire flow curves could be fitted
using the Herschel–Bulkley model according to Eq. (1), as suggested
by Evans and Haisman (1979) in the case of CWMS.
s ¼ s 0 þ K _g n
Where s is the shear stress (Pa), _g is the shear rate (s 1 ), s0 is the
yield stress (Pa), K is the consistency index (Pa s n ) and n is the flow
behaviour index. The values of the Herschel–Bulkley parameters
(s0, K, n) for the different concentrations are given in Table 2.Itisto
be emphasized that the equation could be applied for the whole
curve in the case of CWMS, while it was only valid within the shear
rate range 0.01–2 s 1 for the mixtures as explained in the following.
The steady state flow curves of the mixtures exhibited a rather
different pattern from the ones of CWMS alone. The shape of the
curves was similar to that of CWMS within the shear rate range 2–
300 s 1 but was dramatically shifted to lower shear stress values in
the low shear rate range (10 2 –2 s 1 ). Therefore, the apparent
viscosity of the mixtures was much lower at low shear rates (10 2 –
2s 1 ) than the one of the starch CWMS suspension. The reverse
was seen at high shear rate (2–300 s 1 ) particularly in Figs. 4 and 6,
the apparent viscosity being higher for the mixture than for CWMS
alone. As WPI concentration increased (from 0.5 to 1.5%), the
Table 1
Peaks (mm) of the particle size distribution for WPI, CWMS suspensions and their
mixtures pasted at 90 C and 110 C.
Cooking temperature ( C) CWMS WPI
0% 0.5% 1.0% 1.5%
90 0% – – – 0.3 65 220
3% 48 56 56 56 260
4% 48 65 88 88 301
110 0% – – – 220
3% 56 _ _ 56 410
– not determined.
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686 1681
(1)
shear stress (Pa)
10 2
10 1
10 0
10 -1
10 -2
10 -1
10 0
decrease of apparent viscosity at low shear rate was even more
pronounced. It is also evident that at low shear rate, the flow curve
tends to a plateau suggesting a yield stress. However, from the
shape of these flow curves, it is obvious that the Herschel–Bulkley
equation is not obeyed over the entire shear rate range in the case
of the mixtures. Whatsoever, it was possible to apply this equation
over the range 10 2 –2 s 1 , allowing one to estimate the extrapolated
yield stress of the different suspensions. The Herschel–
Buckley parameters obtained in these conditions are also given in
Table 2. The yield stress decreased from 2.57 Pa (3% CWMS) to
10 1
shear rate (s -1 )
Fig. 4. Equilibrium flow curves plotted in logarithmic scales of 3% CWMS alone (A),
0.5% WPI/3% CWMS (,) and 1.5% WPI/3% CWMS (D) mixtures pasted at 90 C.
Measurement temperature 60 C.
shear stress (Pa)
10 3
10 2
10 1
10 0
10 -2
10 -1
10 0
10 1
shear rate (s -1 )
Fig. 5. Equilibrium flow curves plotted in logarithmic scales of 4% CWMS alone (A),
0.5% WPI/4% CWMS (,) and 1.5% WPI/4% CWMS (D) mixtures pasted at 90 C.
Measurement temperature 60 C.
10 2
10 2
10 3
10 3

1682
shear stress (Pa)
10 2
10 1
10 0
10 -1
10 -2
10 -1
10 0
0.68 Pa (0.5% WPI/3% CWMS) and 0.44 Pa (1.5% WPI/3% CWMS)
when heated at 90 C. In case of 4% CWMS, the yield stress was
dramatically lowered from 11.7 Pa for 4% CWMS alone to 1.2 Pa and
0.7 Pa with 0.5% WPI and 1.5% WPI, respectively. This parameter
decreased from 2.6 Pa (3% CWMS) down to 0.24 Pa (1.5% WPI/3%
CWMS) when the dispersions were heated at 110 C.
From these overall results, it is clear that adding WPI (0.5 to
1.5%) to starch suspensions dramatically modified the flow behaviour
with a marked decrease of the yield stress whatever the CWMS
concentration (3 or 4%) and pasting temperature (90 or 110 C).
Increasing the WPI concentration from 0.5 to 1.5% lowered the yield
stress slightly more. These dramatic modifications in the flow
behaviour at low shear rate reflect important structural differences
between the CWMS suspensions and the WPI/CWMS mixtures.
In Figs. 7–9, the viscoelastic behaviour of these different
suspensions is illustrated by means of the mechanical spectra (G 0
and G 00 as a function of frequency) measured at 60 C. The starch
suspensions (3% or 4%) heated to 90 Cor110 C exhibited a solidlike
behaviour with G 0 > G 00 and G 0 almost independent of
frequency. This is typical behaviour of cross-linked starch suspensions
which is to be ascribed to close-packing of swollen starch
granules (Nayouf et al., 2003; Tecante & Doublier, 1999). The mere
effect of increasing starch concentration from 3% to 4% has been to
increase G 0 , the storage modulus, from 8 to 80 Pa when pasting at
10 1
shear rate (s -1 )
Fig. 6. Equilibrium flow curves plotted in logarithmic scales of 3% CWMS alone pasted
at 90 C(>) and 110 C(A) and 1.5% WPI/3% CWMS pasted at 90 C(6) and 110 C
(:). Measurement temperature 60 C.
Table 2
Herschel–Bulkley constants (s0, K, n) for starch pastes and the mixtures.
Pasting
temperature ( C)
CWMS % (w/w) WPI % (w/w) _g(s 1 ) s0 (Pa) K n
90 3 0 0.01–300 2.57 0.85 0.51
0.5 0.01–2 0.68 1.93 0.55
1.5 0.01–2 0.44 1.66 0.68
4 0 0.01–300 11.75 1.60 0.71
0.5 0.01–2 1.19 8.85 0.49
1.5 0.01–2 0.70 8.15 0.52
110 3 0 0.01–300 2.60 0.92 0.81
1.5 0.01–2 0.24 3.40 0.45
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686
10 2
10 3
G', G'' (Pa)
10 2
10 1
10 0
10 -1
10 0
90 C(Figs. 7 and 9). This is related to the increase of the volume
fraction occupied by swollen starch granules which is the determining
parameter of the rheological properties at this pasting
temperature. Increasing the temperature from 90 Cto110 C (at
3% CWMS) enhanced the G 0 values from 8 to about 20 Pa (Fig. 9).
This is also to be ascribed to an increase of the volume fraction
occupied by swollen starch granules, as indicated by the increase of
the median diameter which varied from 48 to 56 mm (Table 1).
Adding WPI to the starch suspension yielded a dramatic change
in the viscoelastic properties of the system. This is clearly illustrated
for 1.5 WPI/3% CWMS whatever the pasting temperature
(90 Cor110 C) (Figs. 7 and 9) and also for 1.5 WPI/4% CWMS
pasted at 90 C(Fig. 8). The solid-like behaviour of CWMS alone
tends to disappear. Instead, a liquid-like behaviour with
a pronounced dependency of G 0 and G 00 as a function of frequency,
G 0 < G 00 at low frequency and a cross-over of G 0 (u) and G 00 (u) which
was between 2 and 4 rad/s. Such mechanical spectra are reminiscent
of those exhibited by macromolecular solutions. For macromolecular
solutions, below the cross-over, the terminal zone is
accessed with theoretical slopes of 2 for G 0 (u) and 1 for G 00 (u) while,
beyond the G 0 (u) and G 00 (u) cross-over, the so-called plateau zone is
accessed. It is noteworthy that slopes of 2 and 1 for G 0 (u) and G 0 (u),
respectively, in the terminal zone are approached in the cases of 1.5
WPI/4% CWMS at 90 C(Fig. 8) and 1.5 WPI/3% CWMS at 110 C
(Fig. 9). On another hand, the plateau zone is described only in its
first part. Such a rather surprising behaviour is to be paralleled with
the changes in the flow properties above-described, particularly
the decrease of the yield stress.
All these results demonstrate that WPI/CWMS mixtures were
much less structured than the starch suspensions. It is to be noticed
that the mixtures were white and granulous while the CWMS
suspensions appeared translucent and more homogeneous.
3.3. Microscopic observations
Fig. 10 shows the photomicrographs (light microscopy) of the
mixtures cooked at 90 C(Fig. 10a–c) and 110 C(Fig. 10d). In these
10 1
angular frequency (rad/s)
Fig. 7. Mechanical spectra of 3% CWMS alone (A>) and 1.5% WPI/3% CWMS (:6)
mixture pasted at 90 C. Solid and empty symbols represent G 0 and G 00 , respectively.
Measurement temperature 60 C.
10 2

G', G'' (Pa)
10 3
10 2
10 1
10 0
10 -1
10 0
images, only the proteins are contrasted since they have been dyed
using fast Green FCF. For the 0.5% WPI/3% CWMS dispersion
(Fig. 10a), the swollen starch granules were seen, their size being
about 50 mm in agreement with the size determination (Fig. 1).
Moreover, voids could be observed between starch granules. No
large protein aggregate could be distinguished but only small dark
spots suggesting the presence of proteins. At a higher protein
concentration (1 and 1.5% WPI; Fig. 10b and c), large aggregates of
proteins appeared whose size was around 80–200 mm in accordance
with the size determination (Fig. 1). For the higher pasting
10 1
angular frequency (rad/s)
Fig. 8. Mechanical spectra of 4% CWMS alone (A>) and 1.5% WPI/4% CWMS (:6)
mixture pasted at 90 C. Solid and empty symbols represent G 0 and G 00 , respectively.
Measurement temperature 60 C.
G', G'' (Pa)
10 2
10 1
10 0
10 -1
10 0
10 1
angular frequency (rad/s)
Fig. 9. Mechanical spectra of 3% CWMS alone pasted at 90 C(A>), and 1.5% WPI/3%
CWMS (:6) mixture pasted at 110 C. Solid and empty symbols represent G 0 and G 00 ,
respectively. Measurement temperature 60 C.
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686 1683
10 2
10 2
temperature (110 C; Fig. 10d), the dark zones indicating the presence
of proteins were more contrasted suggesting the WPI aggregates
were larger. From these observations, the WPI/CWMS
suspensions could be described as a co-existence of protein-rich
zones and starch-rich zones. The repartition of protein aggregates
throughout the starch dispersion varied with WPI concentration. At
0.5%, the protein spots appeared to be randomly distributed over
the starch dispersion (Fig. 10a). At 1.0% and 1.5% WPI (Fig. 10b and
c), clusters of protein aggregates were formed that seemed to be
localised mostly in the intergranular space between the swollen
starch granules although some of the proteins could be localised
onto the starch granules. These effects were accentuated at 110 C
(Fig. 10d), the strands of protein aggregates being intensified partly
imprisoning starch-rich zones.
A clearer insight into the localisation of the whey protein
aggregates is brought by CLSM microscopy. It is to be emphasized
that ANS is used in order to visualize the proteins. Fig. 11 shows the
photomicrographs of the different dispersions pasted at 90 C. At
1.5% WPI without starch (Fig. 11a), small bright spots were seen
indicating the presence of relatively small protein aggregates.
Fig. 11b shows the result obtained with starch alone (3%). Starch
granules were clearly seen with convoluted surfaces as it is usually
described in the case of cross-linked starch using CLSM whatever
the fluorescent probe (Savary, Handschin, Conde-Petit, Cayot, &
Doublier, 2008; van de Velde, van Riel, & Tromp, 2001; van de
Velde, Weinbreck, Edelman, van der Linden, & Tromp, 2002).
However, the fact that ANS became fluorescent in the presence of
CWMS is rather surprising and suggests the existence of hydrophobic
interactions inside starch granules likely related to the
cross-linking agent. For 0.5% WPI/3% CWMS (Fig. 11c and d), starch
granules were seen in black while the protein aggregates appeared
distributed in the intergranular phase (grey zones) and brighter
spots were seen in the vicinity of swollen starch granules. At 1.5%
WPI/3% CWMS (Fig. 11d and f), this distribution appeared more
clearly with larger spots corresponding to large protein aggregates.
Therefore, part of the proteins was apparently adsorbed onto the
surface of starch granules while the other part was located in the
intergranular phase. It was not possible however to evaluate the
relative ratio of the two populations. This interpretation is
corroborated by phase-contrast microscopy as illustrated in Fig. 12
for 0.5% WPI/3% CWMS (Fig. 12a) and 1.5% WPI/3% CWMS (Fig. 12b)
pasted at 90 C. Particularly in Fig. 12b, protein aggregates are
clearly seen onto the surface of the swollen starch granules.
4. Discussion
The WPI/CWMS mixture can be described as a composite
system of swollen starch granules dispersed in a continuous phase
containing whey proteins. Due to the composition of starch (99%
amylopectin) together with cross-linking, no diffusion of starch
macromolecules is expected towards the surrounding phase at
these pasting temperatures (90–110 C). Regarding the starch alone
suspension, the rheological behaviour is governed by the volume
fraction occupied by the swollen starch granules and their
deformability. The volume fraction varies from w0.50 (3%, 90 C) to
0.65–0.70 (4%, 90 C) (Nayouf et al., 2003).
When heating a WPI/CWMS mixture, two concomitant
phenomena are expected to take place in the same temperature
range (60–80 C), namely protein denaturation, on one side, starch
pasting (swelling), on the other side. From the comparison of size
distributions of CWMS alone and WPI/CWMS mixtures, we can
suggest the swelling of starch granules upon pasting is not modified
by the presence of WPI. This result has been also postulated by
Fitzsimons et al. (2008) for WPI/CWMS suspensions. The intergranular
phase of the system is composed of denatured whey

1684
Fig. 10. LM micrographs of 0.5% WPI/3% CWMS; (a), 1.0%WPI/3% CWMS (b), 1.5% WPI/3% CWMS (c), pasted at 90 C and 0.5% WPI/3% CWMS (d) pasted at 110 C.
proteins that form aggregates whose size depends on the pasting
conditions (heating kinetics, shearing rate, final heating temperature)
and on the protein concentration, this later parameter being
itself governed by the available volume. Since swollen starch
granules occupy 50% to 65–70% of the available volume, the protein
concentration is more than twice the nominal concentration.
Therefore, for 1.5% WPI/3–4% CWMS mixture, the protein concentration
is theoretically higher than the critical gelling concentration
(Cc w1% according to Puyol et al., 2001) for WPI. However, from the
microscopic observations, it is clear that the proteins form large
aggregates in the continuous phase but do not result in a continuous
protein network although the WPI concentration is high
enough to yield a gel. This is likely due to the fact that the mixture is
continuously stirred during pasting in the jacketed vessel. The
kinetics of aggregation may be slowed down upon continuous
stirring compared to the static conditions as it has been showed in
the case of a WPI/xanthan mixture (Walkenström, Nielsen, Windhab,
& Hermansson, 1999).
Disruption of the WPI network by the presence of swollen starch
granules has been reported by several authors (Aguilera & Rojas,
1996; Ravindra et al., 2004; Shim & Mulvaney, 2001). Aguilera and
Rojas (1996) observed that the rigidity of the WPI/corn starch gels
(storage modulus) decreased as the starch mass fraction increased
(from 2 to 7% for a total solid content of 10%). They attributed this
effect to the disruptive effect of starch granules in the continuous
WPI matrix and ascribed this to structural incompatibility. Shim
and Mulvaney (2001) also described two separated networks for
WPI and corn starch with diffusion of amylose in the matrix (at pH
7, 95 C, ratio WPI:corn starch 0.5, total solids 15%). The structure of
the mixed system can then be described as a filled gel where the
continuous phase is a network, itself a mixture of amylose and
protein. In the case of WPI/CWMS systems (pH 7, 75 mM NaCl,
pasting temperature 85 C, total solids 5%), Ravindra et al. (2004)
also showed that at starch concentrations higher than 3% (and 2%
WPI) the swollen starch granules become the predominant phase
interfering with the formation of a continuous WPI network and
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686
weakening the structure of the system. It should be noticed that in
these different studies a moderate stirring was applied during the
heating period which would explain why a continuous protein
network cannot be formed (Fitzsimons et al., 2008). According to
these authors, a continuous protein network can be obtained in the
case of WPI/CWMS mixtures only if no shear is applied during the
thermal treatment, the heating process as well as the cooling one.
In the present work, strong stirring was applied (400 rpm) during
the whole preparation process as would occur in an industrial
process. It is therefore expected that the protein network cannot be
formed in the continuous phase. As a result, the WPI/CWMS system
in the present work can be merely described as swollen starch
granules dispersed in a suspension of WPI aggregates. The
continuous phase appears itself discontinuous with protein
aggregates of different sizes as showed by size distribution
measurements and microscopic observations. On another hand,
part of the largest size aggregates appears stuck onto the swollen
starch granules. These structural features can explain the peculiar
rheological properties of these mixed systems. The solid-like
character of CWMS suspensions is well documented in the literature
(Evans & Lips, 1992; Nayouf et al., 2003). This appears beyond
a critical phase volume (Fc) where close-packing is approached
(Fc > 0.50). This obviously does not imply that a continuous
network is obtained but means that the swollen starch granules
interact between each other due to their vicinity. It is also beyond
Fc that a yield stress is displayed in the flow curves (Nayouf et al.,
2003). Both characteristics, solid-like viscoelastic behaviour, on one
side, and existence of a yield stress in flow measurements, on the
other side, reflect the behaviour of suspensions of close-packed
particles. It might be expected that the addition of WPI would give
rise to a reinforcement of the system with a higher G 0 modulus and
a higher yield stress as was reported by Fitzsimons et al. (2008).
Actually, totally different and unusual behaviours have been
observed. Indeed, the apparent viscosity at high shear rate of the
0.5–1.5% WPI/3% CWMS mixture was increased with respect to
CWMS alone. This is the expected effect of increasing the protein

concentration in the continuous phase. In contrast, however, the
yield stress was decreased. This means that interactions between
the swollen starch granules are weakened by the presence of WPI
aggregates in the continuous phase. In the case of the 1.5% WPI/4%
CWMS stronger changes were seen but the apparent viscosity was
lowered whatever the shear rate range. The changes in the viscoelastic
behaviour were far more dramatic since the solid-like
behaviour was lost and a typical liquid-like behaviour was
observed. The viscoelastic spectrum covered part of the terminal
zone (G 0 < G 00 ) towards the low frequencies and the beginning of
the plateau zone (G 0 > G 00 ) at high frequency. Such a behaviour is
well documented for entangled macromolecular systems as well as
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686 1685
Fig. 11. CLSM micrographs of 1.5% WPI (a), 3% CWMS alone (b), 0.5% WPI/3% CWMS (c) and (d), and 1.5% WPI/3% CWMS (e) and (f) pasted at 90 C. Scale bars refer to 50 mm (a, b, c
and e) and 10 mm (d and f).
wormlike micelles originating from surfactants. However, so far as
we are aware, this has not been reported for suspensions of
deformable particles dispersed in a suspension of denser particles.
We suggest that the presence of large WPI particles in contact
with swollen starch granules modifies the friction forces between
them. This might reduce the number of contact zones between
starch granules or facilitate their relative movement under low
shear forces. The overall consequence would be a destabilization of
the packed structure formed by starch granules. This would occur at
3 and 4%, that is beyond Fc. At 4%, the volume fraction of starch
granules is about 0.65–0.70 which means that the swollen starch
granules are tightly packed (Evans & Lips, 1992). It has been showed

1686
Fig. 12. Phase-contrast micrographs of 0.5% WPI/3% CWMS (a) and 1.5% WPI/3% CWMS
(b) pasted at 90 C.
that the deformability of starch granules plays a major role in the
rheological properties. Under these conditions (4% starch), the
presence of protein aggregates of larger size contributes to lower
the apparent viscosity over the entire shear rate range. The extent
of disorder brought about by protein aggregates to the organised
packed structure of starch granules is still more pronounced than
for 3% CWMS mixtures.
5. Conclusions
Under the present experimental conditions, the WPI/CWMS
system can be described as a suspension of swollen starch granules
whose packing is disrupted by the insertion of protein aggregates
formed upon heating. Microscopic observations revealed that
protein aggregates were distributed unevenly in the continuous
phase, the largest ones being located in the vicinity of swollen
starch granules. The size of the protein aggregates increased with
their concentration in the intergranular space. The pasting
temperature resulted in the same effect. It may be possible to
describe the system as a multiple suspension of a continuous water
phase with different types of dispersed materials (starch granules,
protein aggregates and complex particles of starch granules with
adsorbed protein on the surface). These structural features may
explain the peculiar rheological properties of the mixtures. The
heterogeneity of the continuous phase and the presence of the
large protein aggregates close to starch granules might also explain
the well known grainy texture of food products based on WPI/
CWMS systems which may be governed by both the size and the
repartition of protein aggregates around swollen starch granules.
H. Vu Dang et al. / Food Hydrocolloids 23 (2009) 1678–1686
Acknowledgements
Financial support (research fellowship to HVD) from ‘Conseil
Régional des Pays de la Loire’ (France) is gratefully acknowledged.
The authors are grateful to Sylvie Durand from INRA-BIA (Nantes)
for performance of CLSM observations.
References
Aguilera, J. M., & Baffico, P. (1997). Structure–mechanical properties of heat-induced
whey protein/starch gels. Journal of Food Science, 62, 1048–1053.
Aguilera, J. M., & Rojas, E. (1996). Rheological, thermal and microstructural properties
of whey protein–cassava starch gels. Journal of Food Science, 61, 962–966.
Alloncle, M., Lefebvre, J., Llamas, G., & Doublier, J.-L. (1989). A rheological characterization
of cereal starch/galactomannans mixtures. Cereal Chemistry, 66, 90–93.
Barnes, H. A., Hutton, J. F., & Walters, K. (1989). An introduction to rheology.
Amsterdam: Elsevier Applied Science.
Bourriot, S., Garnier, C., & Doublier, J.-L. (1999). Phase separation, rheology and microstructure
of micellar casein-guar gum mixtures. Food Hydrocolloids, 13, 43–49.
Boye, J. I., Alli, I., & Ismail, A. A. (1996). Effects of physicochemical factors on the
secondary structure of b-lactoglobulin. Journal of Dairy Research, 63, 97–109.
Bryant, C. M., & Mc Clements, D. J. (1998). Molecular basis of protein functionality
with special consideration of cold-set gels derived from heat-denatured whey.
Trends in Food Science & Technology, 9, 143–151.
Durand, D., Gimel, J. C., & Nicolai, T. (2002). Aggregation, gelation and
phase separation of heat denatured globular proteins. Physica Acta, 304, 253–265.
Evans, I. D., & Haisman, D. R. (1979). Rheology of gelatinised starch suspensions.
Journal of Texture Studies, 17, 253–265.
Evans, I. D., & Lips, A. (1992). Viscoelasticity of gelatinized starch dispersions.
Journal of Texture Studies, 23, 69–86.
Fitzsimons, S., Mulvihill, D. M., & Morris, E. R. (2008). Co-gels of whey protein
isolate with crosslinked waxy maize starch: analysis of solvent partition and
phase structure by polymer blending laws. Food Hydrocolloids, 22, 468–484.
Genovese, D. B., & Rao, M. A. (2003). Role of starch granule characteristics (volume
fraction, rigidity and fractal dimension) on the rheology of starch dispersions
with and without amylose. Cereal Chemistry, 80, 350–355.
Kazmierski, M., & Corredig, M. (2003). Characterization of soluble aggregates from
whey protein isolate. Food Hydrocolloids, 17, 685–692.
Loisel, C., Tecante, A., & Doublier, J.-L. (2002). Rheology and structure of cross-linked
starch dispersions. Zywnosc, 9, 169–180.
Mahmoudi, N. (2007). Impact de la structure sur les propriétés interfaciales d’agrégats
de protéines globulaires du lactosérum. PhD Thesis, Université de Nantes.
Matser, A. M., & Steeneken, P. A. M. (1997). Rheological properties of highly crosslinked
waxy maize starch in aqueous suspensions of skim milk components.
Effect of the concentration of starch and skim milk components. Carbohydrate
Polymers, 32, 297–305.
Mulvihill, D. M., & Kinsella, J. E. (1987). Gelation characteristics of whey proteins
and b-lactoglobulin. Food Technology, 41, 102–111.
Nayouf, M., Loisel, C., & Doublier, J. L. (2003). Effect of thermomechanical treatment
on the rheological properties of crosslinked waxy corn starch. Journal of Food
Engineering, 59, 209–219.
Okechukwu, P. E., & Rao, M. A. (1995). Influence of granule size on viscosity of
cornstarch suspension. Journal of Texture Studies, 26, 501–516.
Okechukwu, P. E., & Rao, M. A. (1997). Calorimetric and rheological behaviour of cowpea
protein plus starch (cowpea and corn) gels. Food Hydrocolloids, 11, 339–345.
Puyol, P., Perez, M. D., & Horne, D. S. (2001). Heat-induced gelation isolates (WPI):
effect of NaCl and protein concentration. Food Hydrocolloids, 15, 233–237.
Ravindra, P., Genovese, D. B., Foegeding, E. A., & Rao, M. A. (2004). Rheology of
heated mixed whey protein isolate/cross-linked waxy maize starch dispersions.
Food Hydrocolloids, 18, 775–781.
Savary, G., Handschin, S., Conde-Petit, B., Cayot, N., & Doublier, J.-L. (2008). Structure
of polysaccharide–starch composite gels by rheology and confocal laser scanning
microscopy: Effect of the composition and of the preparation procedure.
Food Hydrocolloids, 22, 520–530.
Shim, J., & Mulvaney, J. (2001). Effect of heating temperature, pH, concentration and
starch/whey protein ratio on the viscoelastic properties of corn starch/whey
protein mixed gels. Journal of the Science of Food and Agriculture, 81, 706–717.
Steeneken, P. A. M. (1989). Rheological properties of aqueous suspensions of
swollen starch granules. Carbohydrate Polymers, 11, 23–42.
Tecante, A., & Doublier, J.-L. (1999). Steady flow and viscoelastic behavior of
crosslinked waxy corn starch kappa-carrageenan pastes and gels. Carbohydrate
Polymers, 40, 221–231.
van de Velde, F., van Riel, J., & Tromp, R. H. (2001). Visualisation of starch granules
morphologies using confocal scanning laser microscopy (CLSM). Journal of the
Science of Food and Agriculture, 82, 1528–1536.
van de Velde, F., Weinbreck, F., Edelman, M. W., van der Linden, E., & Tromp, R. H.
(2002). Visualisation of biopolymer mixtures using confocal scanning laser
microscopy (CLSM) and covalent labelling techniques. Colloids and Surfaces
B-Biointerfaces, 31, 159–168.
Walkenström, P., Nielsen, M., Windhab, E., & Hermansson, A. M. (1999). Effects of
flow behaviour on the aggregation of whey protein suspensions, pure or mixed
with xanthan. Journal of Food Engineering, 42, 15–26.