You are on page 1of 14

Food Hydrocolloids 13 (1999) 303–316

www.elsevier.com/locate/foodhyd

Composite structure formation in whey protein stabilized O/W emulsions.


I. Influence of the dispersed phase on viscoelastic properties
C.K. Reiffers-Magnani a, J.L. Cuq b, H.J. Watzke a,*
a
Nestec Ltd., Nestlé Research Center, Department of Food Science and Process Research, Vers-chez-les-Blanc, P.O. Box 44, CH-1000 Lausanne 26,
Switzerland
b
Université des Sciences et Techniques du Languedoc, Département de Génie Biologique et Sciences des Aliments, Place Eugène Bataillon,
34085 Montpellier Cedex 5, France
Received 10 June 1998; accepted 1 March 1999

Abstract
The viscoelastic properties of whey protein stabilized emulsions containing varying amounts of protein and oil were investigated during
heat treatment using oscillatory rheology. The emulsions were thoroughly characterized prior to gelation by measuring droplet size and
protein adsorption. This emulsion characterization permitted the concentration of protein remaining in the bulk phase of the emulsion to be
determined, by taking into account both oil exclusion effect and protein adsorption, and to prepare a cream and a whey protein solution
corresponding to the emulsion’s dispersed and bulk phases, respectively. Gelation of the phases was investigated separately and was used to
better understand how the oil droplets and bulk proteins contribute to gelation of the whole emulsion. The rheology of polymeric solutions,
giving the gel point as the G 0 – G 00 cross-over, could be applied to the whey protein solution during heat treatment. Regarding the cream, its
gelation revealed that the proteins adsorbed at the O/W interface gelled and were responsible for droplet–droplet interactions, leading to a
particle gelled network. The rheological behavior of the whole emulsion during heat treatment was in-between those of a polymeric gel and a
particle gel. The main parameters influencing the rheology were the bulk protein concentration and the aggregation state of the oil droplets.
The bulk protein concentration determined the formation of a polymeric network in the aqueous phase of the emulsion, and the aggregation
state of the droplets influenced their possible interactions and the formation of a particle network. q 1999 Elsevier Science Ltd. All rights
reserved.
Keywords: Whey proteins; Emulsion; Gelation; Composite; Polymer; Rheology

1. Introduction Shimuzi & Kamiya, 1980; Dickinson, Rolfe & Dalgleish,


1989; Klemaszewski, Das & Kinsella, 1992; Suttiprasit, Al-
A wide range of food products exist as emulsions, either Malah & McGuire, 1993) and gelling (Gault & Fauquant,
liquid or gelled, for instance dairy products, mayonnaise and 1992; Matsudomi, Oshita, Sasaki & Kobayashi, 1992; Hines
sauces. The understanding of structure formation is a key & Foegeding, 1993; McSwiney, Singh & Campanella,
issue in their production because it allows the creation of 1994; Boye, Alli & Ismail, 1996) have been reviewed.
desired textures. They adsorb at the O/W interface and contribute to the
In this context, gellable whey protein stabilized emul- formation of stable emulsions, due to lowering of the inter-
sions are especially suitable models to investigate structure facial tension and to their almost irreversible adsorption
formation. The work presented here concentrates on the (Tornberg, 1978).
importance of the dispersed phase in structure formation Their gelling properties come from their ability to dena-
of a whey protein stabilized emulsion during heat treatment, ture and unfold during heat treatment. Intermolecular inter-
focusing on the importance of the adsorbed protein layer at actions are favored over intramolecular interactions and this
the oil–water interface. can result in formation of a network. The mechanisms
Whey protein isolates exhibit interesting functional prop- involved in protein gelation have been extensively studied
erties for the food industry. Their emulsifying (Yamauchi, (Hermansson, 1979; Steventon, Gladden & Fryer, 1991;
Aguilera, 1995; Hoffmann, van Mil & de Kruif, 1995), espe-
* Corresponding author. Tel.: 1 41-21-785-8026; fax: 1 41-21-785-
cially in terms of the kinetics (Dannenberg & Kessler, 1988;
8554 Taylor & Fryer, 1993) and gel structure (Stading, 1993;
E-mail address: heribert.watzke@rdls.nestle.com (H.J. Watzke) Harwalkar & Kalab, 1985).
0268-005X/99/$ - see front matter q 1999 Elsevier Science Ltd. All rights reserved.
PII: S0268-005 X( 99)00 013-2
304 C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316

Table 1 When other emulsifiers like Tween or lecithin are used, no


Whey protein isolate composition interaction occurs between droplets and matrix and no gel
g/100 g powder strengthening is observed (Dickinson & Yamamoto, 1996;
Matsumara, Kang, Sakamoto, Motoki & Mori, 1993;
Total proteins 89.3 McClements, Monahan & Kinsella, 1993).
Soluble proteins
The work presented here contributes to a better under-
(Total nitrogen × 6.38) 82.6
a -lactalbumin (% total protein) 15.0 standing of the viscoelastic properties of food composites.
b -lactoglobulin (% total protein) 79.0 By studying separately gelation of the dispersed and the
Serum albumin (% total protein) 5.0 bulk phases of a whey protein stabilized emulsion, the
Fat 0.6 contribution of both bulk proteins and oil droplets to gela-
Lactose 0.4
tion of the whole emulsion is evaluated. In particular, gela-
Ash 1.9
Na 1(mg/100 g) 500 tion of the adsorbed proteins is studied in detail and explains
Ca 21(mg/100 g) 150 the particle gel structure of the heated emulsion’s dispersed
K 1(mg/100 g) 100 phase.
Mg 21(mg/100 g) 25
Humidity 6.0
Sum known 98.2 2. Materials and methods

2.1. Materials
The development of viscoelastic properties, which occurs
during gel formation, can be detected by oscillatory rheol- A commercial whey protein isolate (Le Sueur Isolates, Le
ogy. The sol–gel transition is of great importance because it Sueur, MN, USA) was used as the protein source, without
indicates a change in structure. Various methods to detect further purification. The composition of the protein isolate is
this point have been published (Farris & Lee, 1983; given in Table 1. The pH of the powder in deionized water
Apicella, Masi & Nicolais, 1984; Winter & Chambon, was between 7.1 and 7.3 depending on concentration. Soya
1986; Chambon & Winter, 1987; Winter, 1987) and a rheo- oil was used as received from Morgia (Switzerland). Its
logical method, derived from synthetic polymer science and density was 915 kg m 23 at 208C. Tween 40 was purchased
which takes the cross-over of elastic and viscous moduli as from Fluka (Switzerland) and was used as received.
the gel point, is well accepted (Winter & Chambon, 1986;
Chambon & Winter, 1987; Winter, 1987). At this point G 0 ˆ 2.2. Preparation of whey protein solutions
G 00 and the phase angle d is 458.
Whey protein stabilized emulsions gel during heat treat- The whey protein isolate was dissolved in deionized
ment under suitable conditions. A material science approach water at room temperature under moderate magnetic agita-
can be used to study the mechanical properties of the gel tion, to avoid foaming, for 1–2 h. The pH was adjusted to
formed. The gelled emulsion has been described as a 7.0 with 1 N HCl.
composite material where oil droplets play the role of filler Final protein concentrations were 11.2–16.3% w/w.
particles and free proteins form the gel matrix (Langley,
1990). The classical approach to study these emulsions 2.3. Preparation of emulsions
involves gelation of the whole emulsion under either a
Emulsions were prepared by mixing the protein solution
large (Aguilera & Kessler, 1988, 1989; Langley & Green,
or the Tween 40 solution with the oil under vigourous
1989; Aguilera & Kinsella, 1991; Xiong, Aguilera &
magnetic agitation. The final concentrations were 20–
Kinsella, 1991) or small deformation (Van Vliet, 1988;
40% w/w of oil and 9.8–13% w/w of protein or 2% w/w
Aguilera, Xiong & Kinsella, 1993; Dickinson & Yamamoto,
of Tween 40. The protein ensured the formation of a stable
1996). Oil droplets are found to modify the mechanical
emulsion (Magnani-Reiffers, 1997; Hunt & Dalgleish,
properties of the protein gel. They reinforce it by increasing
1994). Both the aqueous solution and the oil were preheated
the bulk protein concentration due to excluded volume;
to 308C. The mixtures were then prehomogenized in a
consequently the proteins strengthen the gel in the bulk
laboratory mixer (Warring Blender, Ismatec SA., Switzer-
phase of the emulsion (Dickinson & Hong, 1995). They
land) for 15 s and immediately homogenized by two
also modify the viscoelastic properties of the gel, by provid-
passages through a Microfluidizer with a head pressure of
ing an interface capable of interacting with the matrix (Van
800 bars (Model 110 T, Microfluidics Corporation, MA,
Vliet, 1988). These interactions depend on the emulsifier
USA).
adsorbed at the interface, and also on the interfacial area
All the emulsions were used within two days of their
which in turn determines the number of connections
preparation.
between the filler and the matrix (Xiong et al., 1991).
When proteins are adsorbed, droplets are well incorporated 2.4. Preparation of creams
within the matrix and the adsorbed proteins form a colloidal
network around them (Jost, Dannenberg & Rosset, 1989). Creams (fat globules and adsorbed emulsifier, i.e. whey
C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316 305

proteins or Tween 40) were obtained after centrifugation of the corrected bulk protein MPB(A) was given by Eq. (3),
20% w/w oil emulsions containing either 10% w/w protein
100
or 2% w/w Tween 40, according to the procedure of Hunt MPB…A† ˆ MP × …3†
100 2 M0
and Dalgleish (1994). The emulsions were centrifuged at
20 000 g and 208C for 1 h. The cream formed was removed with MP the total mass of protein in the emulsion.
and washed with deionized water. The optimal number of In the latter case (B), the corrected bulk protein was
washing cycles was determined by observing changes in the calculated from the quantity of protein (g) adsorbed per
average particle diameter and the presence of free oil on the gram oil kmass determined by the Kjeldahl method (Associa-
top of the cream layer due to coalescence of the droplets. tion of Analytical Chemists, 1980). To validate the calcula-
After three cycles the particle size increased drastically and tions, the protein recovery was evaluated. The sum of the
free oil appeared, therefore, two washings of the cream were protein measured in the cream and in the serum was
adopted as the most appropriate procedure. The whey compared to the original protein in the emulsion and the
protein and Tween 40 stabilized creams contained 66% w/ recovery was 106%. This value was in the range reported
w oil and 2.3% w/w protein for the former and 74% w/w oil previously, between 96–108% (Hunt & Dalgleish, 1994),
for the latter. which ensured that adsorbed proteins are not substantially
For some experiments, an aliquot of the whey protein removed from the interface during the washing procedure.
stabilized cream was taken and split into two parts, which The mass of oil in the cream was calculated by determining
were diluted with deionized water. The first contained the dry mass at 1358C (Mettler PM100 balance, coupled
60% w/w oil and 1.9% w/w protein and the second with a Mettler LP16 desiccator, Mettler Toledo AG.,
contained 33% w/w oil and 1.1% w/w protein. / Switzerland) and correcting for the adsorbed protein
mass.The protein mass adsorbed in the emulsion MPA
2.5. Particle size measurements could thus be calculated:
Particle size in the emulsions and in the creams was MPA ˆ Kmass × M0 : …4†
routinely checked by Photon Correlation Spectroscopy
(PCS 100, Malvern Instruments, England). This method The mass of unadsorbed protein in the emulsion MPF was
has already been used to determine particle size in micro- determined from the following equation:
fluidized systems (Pouliot, Paquin, Robin & Giasson, 1991; MPF ˆ MP 2 MPA : …5†
Robin, Kalab, Britten & Paquin, 1996) and its principle was
previously reported (Schurtenberger & Newman, 1993; To obtain the protein concentration in the bulk phase of
Ricka, 1995). Measurements were performed at a scattering the emulsion, the value obtained from Eq. (5) was corrected
angle of 908 at 258C. The sample was dispersed in ultrapure by the mass of the dispersed phase MDP as follows:
water. The viscosity of the dispersing medium was MDP ˆ M0 1 MPA : …6†
0.89 mPa s and the refractive index of the oil was 1.33.
The determination of the hydrodynamic diameter The corrected protein mass in the aqueous phase MPB(B) of
permitted to calculate the interfacial area AT and the number the emulsion was obtained from Eqs. (5) and (6) as
of fat droplets per 100 g of emulsion NT, according to Eqs 100
(1) and (2) MPB…B† ˆ MPF × : …7†
100 2 MDP
M0 3
AT ˆ × …1†
r0 R0 2.7. Oscillatory rheology
M0 3 Measurements were performed with an oscillatory
NT ˆ × …2†
r0 4pR30 rheometer (Carri-Med CSL 100, TA Instruments, Germany)
fitted with cone and plate geometry (cone diameter 6 cm,
with M0 being the oil mass in the emulsion, r 0 the oil density
angle 38 59 min, gap 107 mm). The samples were poured
and R0, the average hydrodynamic diameter measured in the
onto the plate of the rheometer at 258C and the edge covered
emulsion.
with paraffin oil to prevent dehydration. Whey protein solu-
2.6. Protein content in the emulsion bulk phase tions, emulsions and creams were heated directly on the
rheometer plate, using an integrated Peltier temperature
The protein concentration in the bulk phase of the emul- regulation system. The typical heating protocol was from
sion was calculated by taking into account (A) the oil exclu- 25 to 808C at 1 or 5.58C/min, holding at 808C for 15 min,
sion effect and (B) both the oil exclusion effect and protein cooling to 258C at 5.58C/min and then holding at 258C for
adsorption. All products were prepared by weighing the 15 min. The measurements were performed at a frequency
appropriate mass of oil, protein, and water, so that the of 1 Hz and a strain amplitude of 0.01. In the case of creams,
following calculations were based on mass values, and this strain value was found to damage the structure and so a
expressed for 100 g of emulsion. In the former case (A), strain of 0.005 was used. The geometry chosen allowed a
306 C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316

(a) 1000 0.08


2*Rh
kmass
800
0.06

600
0.04
400

0.02
200

0 0
0 2 4 6 8 10 12 14
Proteins (% w/w)

(b)
16

14

12

10

0
0 2 4 6 8 10 12 14
Total proteins (% w/w)
Fig. 1. Characteristics of a whey protein stabilized emulsion containing 20% w/w, oil and increasing protein content at pH 7.0. (a) Average hydrodynamic
diameter (2 × Rh) and protein/oil ratio (kmass). (b) Bulk protein concentration. no correction (—); correction for the oil exclusion effect (–-); correction for the
oil exclusion effect and the protein adsorption (…).

fast heat transfer from the plate to the sample, even at a the ratio between viscous and elastic moduli, was also
heating rate of 5.58C/min. The gap size fulfilled the used to evaluate the material’s viscoelasticity. The
requirements of a ratio of at least 10 between the gap precision of the gel temperature was ^0.58C and
and the maximum particle diameter (Roberts, 1996). ^2.58C, respectively for heating rates of 1 and 5.58C/
Elastic (G 0 ) and viscous (G 00 ) moduli were followed as min, due to the temperature regulation of the Peltier
indicators of viscoelastic properties. Tan d , obtained as system (Magnani-Reiffers, 1997).
C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316 307

Table 2
Characteristics of whey protein stabilized emulsions at pH 7.0. The total protein content in the emulsions was 9.8% w/w

Oil (% w/w) D (nm) a Nb Ac kmass d Protein in bulk phase e (% w/w)

A B

20 196 6 × 10 21 6.7 × 10 5 0.048 12.3 11.2


30 223 6 × 10 21 8.8 × 10 5 0.059 14.0 11.8
40 300 3 × 10 21 8.7 × 10 5 0.081 16.3 11.6
a
D: average hydrodynamic diameter of the fat globules.
b
N: number of fat globules per 100 g emulsion.
c
A: interfacial area per 100 g emulsion (m 2).
d
kmass: gram protein adsorbed per gram oil.
e
A: the concentration is corrected for the oil exclusion effect. B: The concentration is corrected for the oil exclusion effect and the protein adsorption.

2.8. Transmission electron microscopy size (Fig. 1(a)). The protein adsorption increased as the
diameter decreased. The increased size of the droplets at
The samples were encapsulated in agar gel tubes (Allan- low protein concentrations was due to droplet recoalescence
Wojtas & Kalab, 1984; Kalab, 1988). They were fixed in resulting from the insufficient protein coverage of the origi-
2.5% glutaraldehyde in phosphate buffer. This procedure nal surface area generated (Hunt & Dalgleish, 1994). Above
was followed by fixation in 2% osmium tetroxide (Geis- 2% w/w the protein content was sufficient to ensure a
singer & Stanley, 1981). Dehydratation was performed in complete coverage of the interface, leading to an accumula-
a graded alcohol series starting with 30% alcohol. Spurr tion of the proteins in the bulk phase of the emulsion.
resin was used to embed the samples. Ultra-thin sections The bulk protein concentration of the emulsions, calcu-
(60–80 nm) were stained with uranyl-lead citrate and exam- lated from the quantity of adsorbed proteins and the oil
ined by transmission electron microscopy (Philips CM12, exclusion effect, is depicted in Fig. 1(b). For comparison,
The Netherlands) at an acceleration voltage of 60 kV. the total protein concentration in the emulsion (full line) and
the bulk protein content corrected only for the oil exclusion
effect (dashed line) are also plotted.
3. Results and discussion The bulk protein content depended on the total protein
concentration, but also on the quantity of protein adsorbed at
3.1. Emulsion characterization the O/W interface and on the oil exclusion effect. Below
4.5% w/w, total protein, the bulk protein content corrected
Whey protein stabilized emulsions consist of oil droplets for the oil exclusion effect and the protein adsorption was
stabilized by whey proteins, dispersed in a bulk continuous lower than the total protein concentration, suggesting that
phase containing non-adsorbed proteins. The size of the oil protein adsorption mainly determined the bulk protein
droplets depends on the homogeneization conditions and concentration. The ratio adsorbed protein/unadsorbed
also on the quantity of protein used to prepare the emulsion, protein was high under these conditions. At 4.5% w/w
which determines also the protein adsorption at the O/W protein, the effect of protein adsorption and oil exclusion
interface (Hunt & Dalgleish, 1994) and consequently, the effect canceled out each other. Above this value, the bulk
remaining unadsorbed proteins in the continuous phase of protein concentration was mainly determined by the oil
the emulsion. Gelation of a whey protein stabilized emul- exclusion effect; the ratio adsorbed protein/unadsorbed
sion under appropriate conditions leads to the formation of a protein decreased. In the protein concentration range
composite gel in which the dispersed and the continuous where whey protein can gel, typically around 11% w/w
phases influence the viscoelastic properties (Dickinson & (Fernandes, 1994), the oil content determined the bulk
Hong, 1995; Jost et al., 1989). The microstructural features protein concentration, which is an important result for the
of these emulsions, i.e. particle size, protein adsorption and following emulsion gelation study.
partitioning, were therefore investigated to understand Emulsions containing a constant protein concentration
better how the dispersed droplets and the bulk protein solu- (9.8% w/w) and increasing amounts of oil (from 20 to
tion participate in the gelation process. 40% w/w) were also characterized (Table 2). The average
Fig. 1 depicts the droplet diameter and the quantity of hydrodynamic diameter and the protein/oil ratio increased
protein adsorbed per g of oil of a 20% w/w oil emulsion with oil concentration, which suggests a poor stability of the
containing increasing total protein concentrations. For emulsions with increased oil concentrations, by analogy
protein contents below 1% w/w a sharp decrease of the with Fig. 1(a). Having determined the bulk protein concen-
diameter can be observed, while the curve flattens between tration, an insufficient protein coverage to stabilize the
1 and 2% w/w. Above this concentration, the total protein droplets was improbable as the bulk content was between
concentration in the emulsion has no influence on droplet 11.2 and 11.8% w/w. An explanation may be the increased
308 C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316

100

10

200
1
Gel point 150

100

0.1 50

0
72 73 74 75 76 77 78

Temperature (˚C)

0.01
72 73 74 75 76 77 78
Temperature (˚C)

Fig. 2. Determination of the gel point with the G 0 –G 00 cross-over method for a 13% w/w whey protein solution at pH 7.0. G 0 (black symbols) and G 00 (open
symbols). Variation of phase angle with frequency is plotted in the inserted graph. (X, W) 0.5 Hz; (O, e) 1.0 Hz; (V, S) 5.0 Hz. The heat treatment, performed
in the rheometer, was 18C min 21 from 25 to 808C, at a strain of 0.01.

emulsion viscosity at higher oil contents and, therefore, a and the formation of a network. It indicates that the system
different behavior of the product during emulsification, has undergone transition from a viscous state, where
leading to different emulsion characteristics. The interfacial protein–solvent interactions are favored, to an elastic state
area of the 40% w/w oil emulsion did not increase as where protein–protein interactions take place. The gel point
compared to the 20 and 30% w/w oil emulsions because is of great importance because it indicates transition from a
the higher oil concentration may be compensated by the liquid to a solid state.
increased droplet diameter. The bulk phase protein concen- In the field of synthetic polymers, the rheological
tration was roughly equivalent for the 20, 30 and 40% w/w measurement of the elastic and viscous moduli of a given
oil emulsions, once corrected for the oil exclusion effect and material as a function time or temperature, for example, is
protein adsorption. This result will be considered after- used to determine the gel point. This corresponds to the
wards, when studying emulsion gelation. point where the two moduli cross-over (Winter & Chambon,
1986; Chambon & Winter, 1987; Winter, 1987). Fig. 2
3.2. Gelation of whey protein solutions and emulsions depicts the G 0 and G 00 variations with temperature at various
frequencies for a 13% w/w protein solution. The average
The following study deals with the gelation of whey gelation temperature for a 13% w/w whey protein solution
protein solutions, emulsions and creams. At this point, it was 74.5 ^ 0.38C, which is below the precision of the
is useful to differentiate between polymeric and particle apparatus at a heating rate of 18C/min ( ^ 0.58C). The
gels. Both can be distinguished by their microstructure corresponding phase angle was 45 ^ 28 and does not depend
and rheological properties. The former consists of fine- on the frequency. At pH 7.0 the whey protein solution
stranded molecular networks whose behavior follows the behaved rheologically like a polymeric solution.
synthetic polymer theory (Winter & Chambon, 1986; Fig. 3 shows the effect of emulsion composition on G 0
Chambon & Winter, 1987; Winter, 1987) while the latter and G 00 variations with temperature. Fig. 3(a) depicts the
consists of particle-aggregated networks, whose rheological case for a 20% w/w oil emulsion containing 13% w/w
behavior depends on particle–particle interactions. protein and Fig. 3(b) the case for a 40% w/w oil emulsion
containing 9.8% w/w protein. It was technically not possible
3.2.1. Gel point determination to produce an emulsion with comparable droplet size with
In solution, whey proteins denature and aggregate upon 9.8% w/w protein and 40% w/w oil as with 13% w/w
heat treatment, leading to the formation of an infinite cluster protein and 20% w/w. In the latter, the increased product
at the gel point above the gel concentration threshold. The viscosity led to a substantially higher droplet size after
sol–gel transition corresponds to an increased connectivity emulsification. As droplet size could also influence the
C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316 309

(a) 10000

1000

100

200
10
150
Gel point
1 100

50

0.1 0
68 69 70 71 72 73 74

Temperature (˚C)

0.01
70 71 72 73 74 75
Temperature (˚C)

(b) 1000

100

10

G'
G''
1
25 35 45 55 65
Temperature (˚C)

Fig. 3. Elastic and viscous moduli of whey protein stabilized emulsions at pH 7.0 as a function of temperature. The phase angle is plotted in the inserted graph.
(X, W) 0.5 Hz; (O, e) 1.0 Hz; (V, S) 5.0 Hz. (a) Whey protein emulsion (13% w/w protein and 20% w/w oil). The heat treatment, performed in the rheometer,
was 18C min 21 from 25 to 808C, at a strain of 0.01. (b) Whey protein emulsion (9.8% w/w protein and 40% w/w oil). The heat treatment, performed in the
rheometer, was 5.58C min 21 from 25 to 808C, at an oscillation frequency of 1 Hz and a strain of 0.01.

composite rheological properties, it was, therefore, chosen the protein solution, thus like a polymeric solution.
to produce an emulsion containing less protein but having a However, the gelation temperature was lower for the emul-
comparable droplet size. sion than for the solution, even though both contained the
For the 20% w/w oil emulsion, a G 0 –G 00 cross-over could same total protein concentration.
be determined. The gelation temperature was 70.8 ^ 0.48C, Increasing the oil concentration to 40% w/w led to a
irrespective of the applied frequency. The phase angle was different rheological behavior (Fig. 3(b)); the system had
45 ^ 28 (Fig. 3(a)). The emulsion behaved qualitatively like elastic properties before any heat treatment and no gel
310 C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316

Table 3 3.2.2. Comparison of whey protein solution and emulsion


Composition of emulsions prepared at pH 7.0 gelation parameters
Oil (% w/w) Protein (% w/w) To obtain a better understanding of the influence of the
dispersed phase on the rheological behavior of whey protein
Total Aa Bb stabilized emulsions, three emulsions containing 9.8% w/w
20 9.8 12.3 11.2
30 9.8 14.0 11.8
total protein and increasing amounts of oil from 20 to
40 9.8 16.3 11.6 40% w/w were prepared at pH 7.0 and their gelation was
compared to those of whey protein solutions. The composi-
a
A: the concentration is corrected for the oil exclusion effect. tion of the protein solutions corresponded to the emulsion
b
B: the concentration is corrected for the oil exclusion effect and protein
bulk phase concentrations after correction for (A) the oil
adsorption.
exclusion effect and for (B) the oil exclusion effect and
protein adsorption (Table 3).
Fig. 4 shows the elastic modulus of the emulsions
point could be determined with the classical polymeric (composites) and of the protein solutions corresponding
approach (G 0 –G 00 cross-over). Comparing this behavior to the emulsion bulk phases as a function of the oil concen-
with that of the protein solution presented above, and taking tration in the emulsions. A sharp increase of the composite
into account the corrected bulk protein content of elastic modulus was observed with increasing oil concentra-
11.6% w/w (Table 2), the bulk protein may not be respon- tion, the effect appearing markedly for the most concen-
sible for the observed elastic properties before and during trated emulsion.
heat treatment. The increased oil concentration may be Previous studies have already mentioned the reinforce-
rather responsible for the deviation in behavior of the emul- ment of protein gels by the inclusion of oil droplets (Xiong
sion from that of a polymeric solution. However, the bulk et al., 1991; Aguilera et al., 1993; Dickinson & Hong, 1995;
protein concentration (11.6% w/w) was enough to form a Jost et al., 1989). This reinforcement was explained by the
network and it was felt necessary to define a method which oil exclusion effect (Aguilera et al., 1993; Jost et al., 1989),
allowed the detection of the gel point independently of the which led to an increased protein concentration in the bulk
elastic behavior observed. Fig. 3(b) shows that between 45 phase of the emulsion. However, taking into account the oil
and 658C, the curve went through a minimum. As the exclusion effect (matrix A), the protein solutions had a
proteins represented the only gelling components in the higher elastic modulus than the corresponding emulsions
system, the increased viscoelasticity can be attributed to (Fig. 4), suggesting that the oil exclusion effect alone did
their gelation. The gel point was then determined as the not fully explain the reinforcement of the protein gel.
point where G 0 increased significantly. In the case corre- Taking into account both the oil exclusion effect and the
sponding to Fig. 3(b), the gelation temperature was 56 ^ protein adsorption (matrix B), the protein solutions having
28C. an essentially constant protein concentration (Table 3), had

Fig. 4. Elastic moduli of whey protein stabilized emulsion gels (composite) and whey protein gels (matrixes) as a function of oil concentration. The emulsions
were prepared with 9.8% w/w protein at pH 7.0 and they were heated in the rheometer at 18C min 21 from 25 to 808C, held at 808C during 15 min and cooled to
258C at 5.58C min 21 and held at 258C during 15 min. The frequency was 1 Hz and the strain 0.01. The elastic moduli represented on the graph are those
obtained after the mentioned heat treatment. Matrix A: the protein solution concentration corresponds to that in the emulsion bulk phase after correction for the
oil exclusion effect. Matrix B: the protein solution concentration corresponds to that in the emulsion bulk phase after correction for the oil exclusion effect and
protein adsorption.
C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316 311

Table 4
Comparison of gelation temperature and elastic modulus of whey protein stabilized emulsions (composites) and whey protein solutions (matrixes) at pH 7.0.
The samples were heated in the rheometer at 18C min 21 from 25 to 808C. The frequency was 1 Hz and the strain 0.01 (NB: The standard deviations for Tgel and
G 0 gel correspond to the determination deviations of the gel point with the G 0 –G 00 cross-over)

Tgel(8C) a G 0 gel(Pa)

Oil (% w/w) 20 30 40 20 30 40

Emulsion 79.2 ^ 0.5 71.9 ^ 0.5 57.3 ^ 2.0 8.5 ^ 3.1 11.3 ^ 4.1 35.1 ^ 6.2
Matrix A b 78.3 ^ 0.5 72.8 ^ 0.5 70.9 ^ 0.5 5.5 ^ 2.1 5.0 ^ 2.1 16.1 ^ 5.2
Matrix B c 80.0 ^ 0.5 80.0 ^ 0.5 80.0 ^ 0.5 7.2 ^ 2.3 5.4 ^ 2.3 7.5 ^ 2.1
a
Tgel: Gelation temperature.
b
Matrix A: The protein solution concentration corresponds to that in the emulsion bulk phase after correction for the oil exclusion effect.
c
Matrix B: The protein solution concentration corresponds to that in the emulsion bulk phase after correction for the oil exclusion effect and the adsorbed
protein.

a lower and more constant elastic modulus than the corre- 3.2.3. Heating behavior of emulsion dispersed phase
sponding emulsions (Fig. 4), suggesting that protein adsorp- The behavior during heat treatment of the proteins
tion played a role in the reinforcement of the protein gel. adsorbed at the O/W interface was studied further by
These results are consistent with those of Dickinson and using creams prepared from whey or Tween 40 stabilized
Hong (1995), who used b -lactoglobulin stabilized emul- emulsions. The preparation of the creams was described in
sions containing N-tetradecane and b -lactoglobulin solu- the Materials and Methods (Section 2).
tion. The results presented above showed that whey protein solu-
Gel reinforcement with increasing oil content was tions and whey protein stabilized emulsions containing
also explained by an increased number of particles 20% w/w oil could be regarded as classical polymeric solu-
(oil droplets), leading to a higher connectivity between tions. Their gelation was dictated by the bulk proteins which
filler and matrix (Xiong et al., 1991). The calculation of formed a polymeric network. Increasing the oil concentration
the interfacial area in the 20, 30 and 40% w/w oil emul- led to a non-polymeric behavior, due to the interference of the
sions indicated that this area increased until 30% w/w dispersed phase in the gelation process. The whey protein
oil and then decreased again as a function of the oil stabilized creams could be regarded as a particle network
content (Table 2). Consequently, the increase of the where the dispersed phase dictated the network formation.
composite viscoelastic properties observed with the Fig. 5(a) showed the variations of the elastic modulus of a
increased oil concentration could not be fully explained whey protein and a Tween 40 stabilized cream as a function
by an increased filler–matrix connectivity. The major of time during three cycles of heating at 808C and cooling at
difference observed in the three emulsions was in the 258C (called 1, 2 and 3 in Fig. 5(a)). At the beginning of the
protein adsorption, which nearly doubled when the oil experiment, the moduli of both creams were high, around
concentration increased from 20 to 40% w/w. 5000 Pa. The viscous moduli were around 700 and 50 Pa,
The influence of the dispersed phase on the initial respectively, for the whey and Tween 40 stabilized creams
stage of the network formation was also studied for (data not shown), suggesting that the creams were in a gel
the three emulsions. Table 4 gives the gelation tempera- state (G 0 . G 00 ). Fig. 6 shows the arrangement of the oil
ture and the elastic modulus at the gel point for the globules in the whey protein cream. Close packing of the
emulsions and for their corresponding bulk phases. droplets, together with the measured elastic properties,
The results obtained for the gel point parameters suggest hindering of the oil globules in the two creams.
(Table 4) followed the same trends as those observed However, the elasticity of the Tween 40 stabilized creams
for the final viscoelastic properties (Fig. 4): taking into is between two and four times lower than the elasticity of the
account the oil exclusion effect (matrix A), the protein whey protein stabilized cream, irrespective of the applied
solutions had a lower gelation temperature and a higher strain and throughout the whole temperature cycle (Fig.
elastic modulus than the corresponding emulsions 5(a)). As the only difference between the two creams was
(Table 4). Taking into account both the oil exclusion the nature of the emulsifier adsorbed at the O/W interface, it
effect and protein adsorption (matrix B), the protein could be concluded that the emulsifier influenced the heat-
solution had a higher gelation temperature and an ing behavior of the whole cream.
equal elastic modulus than the corresponding emulsions, The Tween 40 stabilized cream also did not respond in
except for the 40% w/w oil emulsion. These results the same way as the whey protein stabilized cream to varia-
would suggest that the adsorbed proteins rather than tions of applied strain. The behavior of the Tween 40 stabi-
the oil exclusion effect influenced the initial stage of lized cream with heat treatment was not influenced by
gel formation, by decreasing the time until the gel strain. An explanation could be breakage of the structure,
point and enhancing the network connectivity. due to free movement of the droplets around each other.
312 C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316

Fig. 5. Viscoelastic behavior of whey protein or Tween 40 stabilized creams (a) and whey protein solution (b) during heat treatment at pH 7.0. The whey
protein stabilized cream contained 66% w/w oil and 2.3% w/w protein. The Tween 40 contained 74% w/w oil and the whey protein solution was prepared with
13% w/w protein. The heat treatment, performed in the rheometer at an oscillation frequency of 1 Hz, is symbolized by the full line on the graph: three
temperature cycles were made (noted 1, 2 and 3 on the graph), each of them from 25 to 808C.

In the case of the whey protein stabilized cream, elastic protein solution behaved similarly to the cream, and the
properties developed during both the heating and the cool- heating behavior of the whey protein cream was, therefore,
ing periods of the first temperature cycle. However the elas- attributed to gelation of the proteins adsorbed at the O/W
tic modulus increase which occurred on cooling was interface. An increase in the elastic properties could be
reversible (cycle 2 and 3 in Fig. 5(a)). To understand further observed during both the heating and the cooling periods
this behavior and to find an explanation, the elastic proper- of the first temperature cycle. This observation is consistent
ties of the whey protein cream were compared to those of a with previous studies dealing with the whey protein heat
whey protein solution (Fig. 5(b)). Qualitatively, the whey gelation (Dickinson & Hong, 1995). Moreover, reversibility
C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316 313

cream containing 33% w/w oil had an elastic behavior, but


the steep decrease in the elastic modulus suggested break-
down of the structure (arrows in Fig. 7(b)).
Flocculation of the diluted cream could explain this beha-
vior. Gelation of the adsorbed proteins in the interior of flocs
could generate the elasticity found in the 33% w/w diluted
cream. It is, however, not possible from the results reported
in this study to say if flocculation took place before or during
protein gelation. It is possible that either the absence of
proteins in the bulk phase of the emulsion or gelation of
the adsorbed proteins promoted droplet flocculation.

3.2.4. Models of whey protein stabilized emulsion gels


From the results obtained in this study, different micro-
structures can be proposed for whey protein stabilized emul-
sion gels. The models presented in Fig. 8 depict the case of
polymeric, mixed and particle composites.
Polymeric composites consist of gelled emulsions in
which the bulk phase proteins dictate the gelation behavior.
The protein content in the bulk phase is above the protein
gelation threshold, typically 11% w/w (case 1 in Fig. 8).
Polymeric composites are not prone to flocculation of
their dispersed phase, either before or during bulk protein
gelation. These proteins form a polymeric network in which
the oil droplets are entrapped.
Mixed composites consist of gelled emulsions in which
Fig. 6. Microphotograph obtained by TEM of a whey protein stabilized both the bulk phase proteins and the dipersed phase proteins
cream at pH 7.0 before heat treatment. (F: fat globule. P: protein).
dictate the gelation behavior. The protein content in the bulk
phase of the emulsion is above the protein gelation threshold
could be observed in the present study; suggesting that the and the oil droplets tend to interact with each other. Inter-
interactions established during the cooling period are rever- actions between oil droplets can be favored by a high oil
sible. concentration which promotes close packing of droplets or
The results reported in this study showed that the heating by flocculation of the droplets prior to or during heat treat-
behavior of the cream depended on the strain applied. With ment. Polymeric gelation of the bulk protein and particle
the lowest strain, namely 0.005, the maximal elastic modu- gelation of the dispersed phase both contribute to the beha-
lus was observed at 258C for the three temperature cycles. vior of the emulsion upon heating and to the viscoelastic
While the highest, namely 0.01, the maximal elastic modu- properties of the final composite (case 2 in Fig. 8). This
lus is observed during the cooling period of the three model was already suggested by van Vliet (Aguilera et al.,
temperature cycles (Fig. 5(a)). These observations suggest 1993), but no experimental work has so far been performed
breakage of the structure at the highest strain. to support it.
Finally, the above observations on the heating behavior of Particular composites are formed in the absence of bulk
the whey protein and Tween 40 stabilized creams permit to protein matrix formation (cases 3 and 4 in Fig. 8). Depend-
explain better the role of interfacial proteins in the gelation ing on the oil concentration or on emulsion flocculation, the
process: during heating, the proteins adsorbed at the inter- formation of a whole particle gel (case 3 in Fig. 8) or the
face form a gelled structure which contributes, together with formation of cohesive flocs (case 4 in Fig. 8) may arise. In
the close packing of the oil droplets, to the formation of a the former, close packing of the droplets permits the
particle network. The sensitivity of the whey protein cream adsorbed proteins to form a network occupying the whole
to the applied strain suggested that the network surrounding volume of the sample (case 3 in Fig. 8). In the latter, gelation
the oil droplets is fragile. of adsorbed proteins around a limited number of oil droplets
The heating behavior of the adsorbed protein was also leads to the formation of dispersed units of gels in the whole
investigated in diluted creams (Fig. 7), which could simulate volume of the sample (case 4 in Fig. 8). In the present study,
a whey protein stabilized emulsion in which the bulk phase this case arise with diluted cream and was attributed to the
contained no proteins. With 60% w/w oil the diluted cream absence of protein in the bulk phase. From a macroscopic
behaved qualitatively like the non-diluted cream, with a G 0 point of view, the sample had elastic properties, as measured
value approximately 10 times lower (Fig. 7(a)). The diluted by rheology, but appeared visually like a liquid.
314 C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316

(a) 1500 100

80

1000
60

40
500

20

0 0
0 20 40 60 80
1 2 3
Time (min)

(b) 15 100

80

10
60

40
5

20

0 0
0 20 40 60 80
1 2 3
Time (min)

Fig. 7. Viscoelastic behavior during heat treatment of diluted whey protein stabilized creams at pH 7.0. (a) The diluted cream contained 60% w/w oil and
1.9% w/w protein. The diluted cream contained 33% w/w oil and 1.1% w/w protein. The heat treatment, performed in the rheometer at an oscillation frequency
of 1 Hz and a strain of 0.005, is symbolized by the full line on the graph: Three temperature cycles were made (noted 1, 2 and 3 on the graph), each of them from
25 to 808C.

4. Conclusion interface and then to calculate the remaining protein


concentration in the emulsion’s bulk phase. Then a cream
The heat gelation of whey protein stabilized emulsions equivalent to the emulsion dispersed phase (whey protein
and their viscoelastic properties after heat treatment were stabilized oil droplets) and a whey protein solution equiva-
investigated. As a first step, we thoroughly characterized the lent to the emulsion bulk phase were prepared and their
emulsions prior to heating. This characterization permitted gelation was investigated.
to determine the quantity of proteins adsorbed at the O/W Bulk phase proteins form a polymeric network while the
Before heat treatment After heat treatment

"Polymeric composite"
1
B ulk p ha se p ro te in a b o ve g e la tio n thre sho ld P o lym e ric n e tw o rk

C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316


O il d ro p le ts d o no t inte ra ct with e a ch o the r
Bulk proteins G e lle d a d so rb e d
p ro te in s
Fat globule

Adsorbed proteins

P a rticle n e tw o rk
2 "Mixed composite"
B ulk p ha se p ro te in a b o ve g e la tio n thre sho ld
O il d ro p le ts inte ra ct with e a ch o the r
P o lym e ric n e tw o rk
(clo se p a cking , flo ccula tio n)

"Particle composite"
3
P a rticle n e tw o rk
B ulk p ha se p ro te in b e lo w g e la tio n thre sho ld
D ro p le t clo se p a cking

4 "Particle composite" F lo c
B ulk p ha se p ro te in b e lo w g e la tio n thre sho ld
D ro p le t flo ccula tio n

315
Fig. 8. Propositions for gelled microstructures in heated whey protein stabilized emulsions at pH 7.0. Cases 1, 2, 3 and 4 are described in the text.
316 C.K. Reiffers-Magnani et al. / Food Hydrocolloids 13 (1999) 303–316

dispersed phase forms a particle network. The viscoelastic Hermansson, A.-M. (1979). In A. Pour-El, Functionality and whey protein
properties of the gelled emulsion after heat treatment were structure, (p. 81–104). Washington, DC: American Chemical Society.
Hines, M. E., & Foegeding, E. A. (1993). J. Agric. Food Chem., 41, 341–
in between those of a whey protein solution and a whey 346.
protein cream, depending on the protein concentration in Hoffmann, M. A. M., van Mil, P. J. J. M., & de Kruif, C. G. (1995).
the bulk phase and the aggregation state of droplets prior Conference Proceedings. In E. Dickinson & D. Lorient, Food
to and during the heat treatment. Close packing or floccula- macromolecules and colloids, (p. 171–177). London: Royal Society
tion of the droplets promoted droplet–droplet interactions of Chemistry.
Hunt, J. A., & Dalgleish, D. G. (1994). Food Hydrocol., 8, 175–187.
and lead to the emulsion’s behavior changing from poly- Jost, R., Dannenberg, F., & Rosset, J. (1989). Food Microstruct., 8, 23–28.
meric to particle gel behavior. Kalab, M. (1988). Food Microstruct., 7, 213–214.
Klemaszewski, J. L., Das, K. P., & Kinsella, J. E. (1992). J. Food. Sci., 57,
366–371.
Acknowledgements Langley, K. R. (1990). Food quality and preference, 2, 111–115.
Langley, K. R., & Green, M. L. (1989). J. Text. Studies, 20, 191–207.
This work is part of a Ph.D Thesis, which was supported Magnani-Reiffers, C.K. (1997). Ph.D Thesis. Laboratoire de Génie Biolo-
by the Nestlé Research Center, Switzerland. The Center is gique et Sciences des Aliments, Université des Sciences et Techniques
de Montpellier, Montpellier, France.
gratefully acknowledged for technical and financial support. Matsudomi, N., Oshita, T., Sasaki, E., & Kobayashi, K. (1992). Biotech.
The authors gratefully acknowledge M.-L. Dillmann and Biochem., 56, 1697–1700.
M.-F. Clerc for their skilled microscopic measurements. Matsumara, Y., Kang, I. J., Sakamoto, H., Motoki, M., & Mori, T. (1993).
Food Hydrocol., 7, 227–240.
McClements, D. J., Monahan, F. J., & Kinsella, J. E. (1993). J. Text.
References Studies., 24, 411–422.
McSwiney, M., Singh, H., & Campanella, O. H. (1994). Food Hydrocol., 8,
Aguilera, J. M. (1995). Food Technol., 10, 83–89. 441–453.
Aguilera, J. M., & Kessler, H. G. (1988). Milchwiss., 43, 411–415. Pouliot, Y., Paquin, P., Robin, O., & Giasson, J. (1991). Int. Dairy J., 1, 39–
Aguilera, J. M., & Kessler, H. G. (1989). J. Food Sci., 54, 1213–1221. 49.
Aguilera, J. M., & Kinsella, J. E. (1991). J. Food Sci., 56, 1124–1128. Ricka, J. (1995). Lecture Notes. ETH Zurich, Switzerland.
Aguilera, J. M., Xiong, Y. L., & Kinsella, J. E. (1993). Food Res. Intern., Roberts, I.D. (1996). PhD Thesis, University of Wales, Swansea, England.
26, 11–17. Robin, O., Kalab, M., Britten, M., & Paquin, P. (1996). Lait, 76, 551–570.
Allan-Wojtas, P., & Kalab, M. (1984). Food Microstruct., 3, 197–198. Schurtenberger, P., & Newman, M. E. (1993). Environmental analytical
Apicella, A., Masi, P., & Nicolais, L. (1984). Rheol. Acta, 23, 291–296. and physical chemistry series. In J. Buffle & H. P. van Leeuwen,
Association of Analytical Chemists (1980). Official methods of analysis Environmental particles, (p. 37–115). London: Lewis.
(13th Ed.). Stading, M. (1993). PhD dissertation, SIK, Chalmers University of Tech-
Boye, J. I., Alli, I., & Ismail, A. A. (1996). J. Agric. Food Chem., 44, 996– nology, Göteborg, Sweden.
1004. Steventon, A. J., Gladden, L. F., & Fryer, P. J. (1991). J. Text. Studies., 22,
Chambon, F., & Winter, H. H. (1987). J. Rheol., 31, 683–697. 201–218.
Dannenberg, F., & Kessler, H. G. (1988). J. Food Sci., 53, 258–263. Suttiprasit, P., Al-Malah, K., & McGuire, J. (1993). Food Hydrocol., 7,
Dickinson, E., & Hong, S.-T. (1995). J. Agric. Food Chem., 43, 2560– 241–253.
2566. Taylor, S. M., & Fryer, P. J. (1993). Food Hydrocol., 6, 543–557.
Dickinson, E., Rolfe, S., & Dalgleish, D. G. (1989). Food Hydrocol., 3, Tornberg, E. (1978). J. Sci. Food Agric., 29, 867–879.
193–203. Van Vliet, T. (1988). Coll. Polym. Sci., 266, 518–524.
Dickinson, E., & Yamamoto, Y. (1996). Food Hydrocol., 10, 301–307. Winter, H. H. (1987). Polym. Eng. Sci., 27, 1698–1702.
Farris, R. J., & Lee, C. (1983). Polym. Eng. Sci., 23, 412–586. Winter, H. H., & Chambon, F. (1986). J. Rheol., 30, 367–382.
Fernandes, P. B. (1994). Food Hydrocol., 8, 277–285. Xiong, Y. l., Aguilera, J. M., & Kinsella, J. E. (1991). J. Food Sci., 56, 920–
Gault, P., & Fauquant, J. (1992). Lait., 72, 491–510. 925.
Geissinger, H. D., & Stanley, D. W. (1981). Scan. Electron Microsc., 3, Yamauchi, K., Shimizu, M., & Kamiya, T. (1980). J. Food Sci., 45, 1237–
415–418. 1242.
Harwalkar, V. R., & Kalab, M. (1985). Milchwiss., 40, 65–68.

You might also like