Food Hydrocolloids 20 (2006) 68–78

www.elsevier.com/locate/foodhyd

b-Glucan prepared from spent brewer’s yeast as a fat

replacer in mayonnaise
Supachai Worrasinchai, Manop Suphantharika*, Surinya Pinjai, Pimon Jamnong
Department of Biotechnology, Faculty of Science, Mahidol University, Rama 6 Road, Bangkok 10400, Thailand

Received 20 July 2004; revised 24 February 2005; accepted 3 March 2005

Abstract
Application of b-glucan prepared from spent brewer’s yeast as a fat replacer in mayonnaise was studied. Fat was partially substituted by
b-glucan at levels of 25, 50, and 75% which were referred to as 25B, 50B, or 75B formulations, respectively. The full fat (FF) (100% oil)
mayonnaise without b-glucan substitution was used as a control experiment. Physicochemical, rheological, and microbiological analyses,
and sensory evaluation of the FF and reduced fat (RF) mayonnaises were performed. The results indicated that all RF mayonnaises had
significantly lower energy content, but higher water content than their FF counterpart and these differences increased with increasing
substitution levels of b-glucan. With regard to pH, there were no significant differences between the FF and RF mayonnaises after one-day
storage. However, pH values of the RF samples decreased with an increasing degree of substitution after 2 months storage. In terms of
texture, the 50B and 75B formulations showed similar firmness and adhesiveness values as those of the FF sample. Microstructure analysis
showed close packing structures of large droplets for the FF and 25B and loose structures of a network of aggregated small droplets for the
50B and 75B samples. Both FF and RF mayonnaises exhibited thixothopic shear thinning behaviour under steady shear tests and were
rheologically classified as weak gels under small amplitude oscillatory shear tests. Microbial loadings of all mayonnaises tested were in the
acceptable limit throughout the storage time. The RF mayonnaises exhibited higher storage stability than the FF sample. Sensory evaluation
demonstrated that mayonnaises substituted with b-glucan of not more than 50% were acceptable. This study shows good potential for spent
brewer’s yeast b-glucan to be used as a fat replacer in mayonnaise.
q 2005 Elsevier Ltd. All rights reserved.

Keywords: b-Glucan; Fat replacer; Mayonnaise; Brewer’s yeast; Emulsion

1. Introduction basic formula shall be replaced by other ingredients. There

Mayonnaise is an oil-in-water emulsion and is a widely
consumed product. Mayonnaise must contain at least 78.5%
total fat and 6% pure egg yolk (Codex Alimentarius
Commission, 1989). Fats have many important functions
on a food emulsion. They contribute to the flavour,
appearance, texture and shelf life of food emulsion in high
specific ways (McClements & Demetriades, 1998). The
apparent relationship between dietary fat and development
of cardiovascular diseases, hypertension and obesity has
prompted a consumer to be more trend to demand reduced
fat products. To produce light mayonnaise, fat from the
* Corresponding author. Tel.: C66 2 201 5314; fax: C66 2 354 7160.
E-mail address: scmsp@mahidol.ac.th (M. Suphantharika).
0268-005X/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodhyd.2005.03.005
is a considerable challenge for preserving the same quality
attributes as the original full fat products.
b-Glucan is described as a polymer of glucose that is
widespread in many bacteria, fungi, mushrooms, algae and
higher plants (such as oat and barley). One important source
of b-glucan is the cell wall of yeast (Saccharomyces
cerevisiae) which contains about 55–65% of b-glucan (Klis,
Mol, Hellingwerf, & Brul, 2002). In this experiment, b-
glucan was extracted from spent brewer’s yeast, a by-
product from a brewery. This b-glucan had high apparent
viscosity, water holding, oil binding, and emulsion stabiliz-
ing capacities (Thammakiti, Suphantharika, Phaesuwan, &
Verduyn, 2004).
b-Glucan is poorly utilised in the human digestion
tract and, therefore, functions as a non-caloric food.
Moreover, it has attracted attention because of its bioactive
and medicinal properties such as immune-stimulating,
anti-inflammatory, anti-microbial, anti-infective, anti-viral,
 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78
anti-tumoral, cholesterol-lowering, radioprotective and
wound-healing (Kogan, 2000; Stone & Clarke, 1992).
There is little published data in the literature on
application of hydrocolloids (gums) to partially replace fat
in mayonnaise. Viscoelastic properties of reduced fat
mayonnaise were improved by xanthan gum addition (Ma
& Barbosa-Cánovas, 1995) and addition of algal alginates to
the full fat (70% oil) mayonnaise as stabilizers resulted in an
improvement of emulsion strength (Mancini, Montanari,
Peressini, & Fantozzi, 2002). Volikakis, Biliaderis, Vam-
vakas, and Zerfiridis (2004) used cereal b-glucan as a fat
replacer in low-fat cheese and thus led to a significant
improvement in texture of the product. However, the colour,
flavour and overall acceptability scores were adversely
effected. More recently, Kontogiorgos, Biliadens, Kiosseo-
glou, and Doxastakis (2004) demonstrated that cereal b-
glucans could be used as stabilizers in model salad
dressings. However, no information has been published on
the application of spent brewer’s yeast b-glucan as a fat
replacer in mayonnaise. The use of this b-glucan prep-
aration to partially substitute vegetable oil in the formu-
lation of these foodstuffs has the two-fold advantage of (1)
decreasing the calorie content of the emulsions, because of a
lower proportion of oil in the formulation and (2) utilizing
products derived from industrial by-products. This type of
mayonnaise is suitable for health conscious people who are
concerned about their fat intake and who are on a diet.
The objectives of this study, therefore, were to
investigate the effects of partial b-glucan substitution for
fat in mayonnaise formulations on physicochemical,
rheological, microbiological, and sensorial properties of
the reduced fat (RF) mayonnaises. The full fat (FF)
formulation was used as reference.
2. Materials and methods
2.1. b-Glucan preparation
b-Glucan was prepared according to the procedure
described in our preceding article (Thammakiti et al.,
2004). Briefly, spent brewer’s yeast slurry was autolysed at
50 8C for 24 h. Yeast cell walls were collected by
centrifugation. These cell walls were homogenized,
extracted with alkali followed by acid and then washed
with distilled water. The obtained b-glucan preparation was
a light-tan coloured paste having a pH value of 4.29 and a
proximate composition in percentage (w/w) as followed:
moisture 93.37, fat 0.07, ash 0.04, protein 0.38, and
carbohydrate 6.13.
2.2. Mayonnaise preparation
One kilogram of each mayonnaise sample was prepared
in this study. Full fat (FF) mayonnaise recipe was modified
from that recommended by Xiong, Xie, and Edmondson
69
(2000). The recipe contained the following ingredients in
percentage (w/w): pure egg yolk 7.27, vinegar (5% (w/v)
acetic acid) 8.72, soybean oil 82.19, salt 0.73, mustard 0.36,
sugar 0.36, and ground white pepper 0.36. All ingredients
used were purchased from a local supermarket. In this
experiment, for the reduced fat (RF) mayonnaises, soybean
oil was replaced by b-glucan at levels of 25% (205.5 g
b-glucan), 50% (401.0 g b-glucan) and 75% (616.5 g
b-glucan) of total oil used and were referred to as the
25B, 50B, and 75B formulations, respectively. The
preparation was as follows; firstly egg and vinegar were
mixed together and then all other ingredients (including
b-glucan for the RF formulations) except oil were added and
stirred homogeneously by a mechanical overhead stirrer
(Ika model 20 RW, Janke and Kunkel GmbH and Co. KG,
Staufen, Germany). The oil was finally poured in very
slowly, while stirring at 1600 rpm for 1 min and followed by
2000 rpm for another 4 min. The FF mayonnaise was
obtained at this stage, whereas the pre-emulsions were
obtained for the RF mayonnaises. These pre-emulsions were
passed through the high-pressure homogenizer. The press-
ures and number of passes were varied according to the
physical and rheological characteristics of each formulation.
The pressures of 100/40, 600/60, and 400/40 bar for the
first/second stages homogenizing values and the number of
passes of 4, 3, and 3 passes were used for the 25B, 50B and
75B formulations, respectively. Subsamples (z200 g) of
mayonnaise were aseptically transferred to sterile 250 ml,
Duran glass bottles with polypropylene screw caps and left
one-day at room temperature (25–30 8C) before testing.
2.3. Proximate analysis
Moisture, protein, and ash contents were determined
according to AOAC (1990) official methods. Fat content
was determined by Bligh and Dyer method (Bligh & Dyer,
1959). Carbohydrates were determined by subtracting the
sum of moisture, protein, fat, and ash percentages from
100%.
2.4. Total plate count
Total plate counts of each sample (1 g) were determined
by serial (1:10) dilution in sterile 0.1% peptone and pour
plating (1 ml) in triplicate on plate count agar (Merck
KGaA, Darmstadt, Germany). Colonies were counted after
the plates were incubated for 48 h at 37 8C.
2.5. Colour measurement
Mayonnaise samples were measured for colour in the L*,
a*, b* system using a Minolta Colorimeter CR-300 (Konica
Minolta Business Technologies, Inc., Langenhagen/Hann-
over, Germany) which was calibrated using a white standard
porcelain plate (L*Z97.10, a*ZK0.07, b*ZC1.97).
A fixed amount of mayonnaise was poured into the
70 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78
measurement cup, which was then surrounded with a black
paper strip. In this colour system, L* represents the
lightness, and a* and b* are colour coordinates: where C
a is the red, Ka is the green, Cb is the yellow, and Kb is
the blue directions.
2.6. pH and water activity (aw) measurements
pH and aw values were measured at a temperature of
25 8C using a Denver pH meter model 225 (Denver
instrument, CO, USA) and a Aqualab water activity meter
(Decagon Devices, Inc., WA), respectively. Three replicate
readings (three different samples) were taken for each pH
and aw measurements of mayonnaises.
2.7. Particle size measurement
Mayonnaise samples (0.04 g) were diluted with 150 ml
0.1% sodium dodecyl sulfate (SDS) solution, stirred
gently by a spatula until completely dispersed. Samples
were analyzed by a particle size analyzer (Mastersizer
2000, Malvern Instrument Ltd, Worestershire, UK). The
sample solution was dispersed in distilled water at
1200 rpm until an obscuration rate of 0.2–0.5% was
obtained. Background and sample were measured for
10 s. Optical properties of the sample were defined as
follow: refractive index 1.460 and absorption 0.00
(Roland, Piel, Delattre, & Evrard, 2003). Droplet size
measurements arePreportedP as the volume-weighted mean
diameter: d4;3 Z ni di4 = ni di3 , where ni is the number
of droplets of diameter di. Each sample was measured in
triplicate.
2.8. Scanning electron microscopy (SEM)
Sample preparation was according to the procedure of
Egelandsdal et al. (2001). The mayonnaise samples were
encapsulated in 2.5% agar at 45 8C, and then cut into 3–
4 mm cubes. The gel samples were fixed overnight in 2%
glutaraldehyde in 0.1 M phosphate buffer (pH 7), rinsed
in several changes of buffer and post fixed overnight in
2% osmium tetraoxide with 0.1 M imidazole. After
washing the samples with buffer, dehydrated in a graded
ethanol series (50, 70, 90, 95, 99 and 100%, 30 min in
each). The samples were critical point dried through CO2
in a Hitachi critical point dryer HCP-2 (Hitachi, Tokyo,
Japan). Dry sections were fractured and torn with a blade
and fragment were mounted on aluminium SEM stubs,
and coated with platinum/palladium. The microscope
used was the Hitachi SEM model s-2500 (Hitachi,
Tokyo, Japan) at 15 kV and a magnification of 2000!.
2.9. Texture profile analysis
Mayonnaise texture measurements were carried out with
the TA.XT2i Texture Analyser (Stable Micro Systems Ltd,
Surrey, UK) with a 5 kg load cell. Back extrusion cell with
35 mm diameter compression disc was used. The samples
were carefully scooped into acrylic cylindrical containers
(50 mm internal diameter and 75 mm height) to a depth of
55 mm. One cycle was applied, at a constant crosshead
velocity of 1 mm/s, to a sample depth of 40 mm, then
returned. From the resulting force–time curve, the values for
texture attributes, i.e. firmness and adhesiveness, were
obtained using the Texture Expert for Window Version 1
equipment software. Firmness was the maximum force as
the test cell penetrated into the sample as described by
Mohamed and Morris (1987). Adhesiveness was the
negative force area representing the work necessary to
pull the compressing plunger away from the sample as
defined by Bourne (1978).
2.10. Viscoelastic behaviour
Rheological measurements were performed with a
rheometer (Physica MCR 150, Anton Paar GmbH, Stutt-
gard, Germany) with a cone and plate geometry sensor
(50 mm diameter, 28 cone angle). The measurements were
conducted at a gap distance of 0.05 mm. All samples were
measured at 25 8C. The linear viscoelastic range was
determined with strain sweep (0.1–100%) at a fixed
frequency of 10 Hz. After that, a dynamic frequency
sweep was conducted by applying a constant strain of
0.5% which was within the linear region, over a frequency
range between 0.1 and 100 Hz.
2.11. Steady flow behaviour
The shear stress compared with shear rate data were
obtained immediately after the frequency sweep data as
described above. They were obtained first as the shear rates
were increased continuously up to 300 sK1 followed by
shear rates in a decreasing mode from 300 to 0 sK1 in 6 min,
the former are referred to as ‘up’ curves the latter as ‘down’
curves. The shear stress–shear rate data were analyzed by
using the power law model as follows
s Z Kgn
where s is shear stress (Pa), g is shear rate (sK1), K is
the consistency coefficient (Pa sn), n is the flow behaviour
index (–).
2.12. Stability test
The mayonnaise samples were stored at room tempera-
ture (25–30 8C) during which instrumental texture profile
analysis, particle size analysis, and microbiological analysis
were performed every week for 2 months.
 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78
2.13. Sensory evaluation
Sensory evaluation was conducted on the mayonnaise
samples after one-day storage at room temperature. Sensory
characteristics: appearance, colour, odour, texture, taste,
and overall acceptability were evaluated by 30 trained panel
on 9-point hedonic scale, 1Zthe least, the lowest; 9Zthe
most, the highest. Bread was provided between samples to
cleanse the palate.
2.14. Statistical analysis
A one-way analysis of variance (ANOVA) and Tukey’s
test (P%0.05) were used to establish the significance of
differences in the physicochemical properties, proximate
analysis and sensory evaluation of mayonnaise samples.
The result was performed using the SPSS version 9.0
windows program (SPSS, Inc., 1998, Chicago, IL).
3. Results and discussion
3.1. Chemical composition and caloric values
The proximate analysis and caloric values of the reduced
fat (RF) mayonnaises containing b-glucan, and the full fat
(FF) mayonnaise (control) are listed in Table 1. The
moisture content increased with increasing levels of
b-glucan substitution due to a very high moisture content
of b-glucan preparation which is a typical characteristic of
carbohydrate-based fat replacers (Akoh, 1998). No signifi-
cant difference between the RF and FF mayonnaises in ash
and protein contents was found. While carbohydrate content
slightly increased with increasing levels of substitution. The
caloric values of the RF formulations were significantly
(P%0.05) reduced with increasing the levels of b-glucan
replacement. Besides water as a main component of
b-glucan sample, the b-glucan itself is non-caloric because
it is not digested or absorbed in the human digestive tract
(Reed & Nagodawithana, 1991).
3.2. Physicochemical and microbiological analyses
The pH, aw, and total plate count values of the FF and RF
mayonnaises after storage for one-day and 64 days at room
Table 1
Proximate analysis (%, w/w) and caloric values of mayonnaise samples
Samples Moisture content Fat Carbohydrate
FF 11.49G0.05d 84.88G0.44a 1.62G0.48b
25B 29.68G0.48c 65.69G1.19b 2.66G0.90ab
50B 50.73G0.24b 43.35G1.39c 3.75G1.47ab
75B 68.98G0.15a 24.16G0.07d 4.59G0.28a
Assays were performed in triplicate. MeanGSD values followed by the same letter in each column are not significant different at P%0.05 by ANOVA and
Tukey’s test.
a
Caloric valuesZ(4!protein)C(9!fat)C(4!carbohydrate). Source: Souci, Fachman, and Kraut (2000).
71
temperature (25–30 8C) are shown in Table 2. There was no
significant (PO0.05) difference in pH values (3.80–3.85) of
all formulations at one-day storage. Since the amount of
water significantly (P%0.05) increased with increasing
percentage of b-glucan substitution, one might expect that
the pH of the RF formulations would be higher than that of
the FF sample due to the dilution of acetic acid in the
aqueous phase of the RF samples (Hathcox, Beuchat, &
Doyle, 1995). However, this was not the case in our study.
There are two possible reasons to explain our findings: (1)
undissociated acetic acid is slightly soluble in oil, therefore,
the pH of mayonnaise would increase as the percentage of
oil increase, particularly after the oil reach 50% level
(Xiong et al., 2000), (2) acetic acid residue remained in the
b-glucan preparation after acid extraction. The constant pH
values observed would result from the balance between the
dilution effect and a compensation of acetic acid from the
b-glucan preparation. From a microbiological safety point
of view, it is generally recommended that mayonnaise made
with unpasteurized egg is prepared with vinegar to a pH of
4.1 or less and held at room temperature (18–22 8C) for at
least 24 h to reduce the risk from microorganisms (Radford
& Board, 1993). The aw of mayonnaises increased, as
expected, with increasing percentage of b-glucan substi-
tution mainly due to the increasing of water content of the
formulations. Chirife, Vigo, Gómez, and Favetto (1989)
reported that the aw of regular mayonnaises (77–79% oil)
was about 0.93 and that of diet samples (37–41% oil) was
higher, e.g. close to 0.95. These authors demonstrated that
the aw of mayonnaise was largely determined by the
chemical nature and concentration of soluble solids in the
aqueous phase. The aw was unchanged during 64 days
storage because the samples were kept in closed containers
and water loss due to evaporation was eliminated.
After one-day storage, the total plate counts of the FF
formulation were significantly (P%0.05) higher than those
of the RF mayonnaises, whereas there is no significant
difference of total plate counts between the RF formu-
lations. This was possibly due to the higher oil content of the
FF formulation gave the higher possibility that microor-
ganisms may reside in the oil phase and do not see the acid
and pH effect in the aqueous phase and hence the levels are
not reduced (Pourkomailian, 2000). After 64 days storage,
the total plate count of the FF sample decreased, whereas
those of the RF ones increased. The reduction of microbial
Protein Ash Caloric valuesa (kcal/100 g)
1.20G0.02a 0.81G0.00ab 775.17G2.07a
1.18G0.15a 0.78G0.00b 606.61G7.75b
1.33G0.05a 0.84G0.02a 410.48G6.62c
1.43G0.11a 0.83G0.02a 241.56G0.44d
72 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78

Table 2
pH, water activity, and microbiological analysis of mayonnaise samples

Samples pH Water activity (aw) Total plate count (cfu/g)

One-day 64 days One-day 64 days One-day 64 days
FF 3.84G0.03a 3.91G0.02a 0.958G0.003c 0.943G0.006c 277G85.0a 77G32.2b
25B 3.80G0.02a 3.52G0.01b 0.989G0.003b 0.984G0.002b 63G25.2b 123G20.8b
50B 3.84G0.02a 3.44G0.01c 0.995G0.001a 0.994G0.001a 100G81.9ab 447G49.3a
75B 3.85G0.01a 3.33G0.01d 0.998G0.001a 0.999G0.001a 127G106.9ab 390G75.5a

Assays were performed in triplicate. MeanGSD values followed by the same letter in each column are not significant different at P%0.05 by ANOVA and
Tukey’s test.

loadings of the FF sample was probably due to the effect of
undissociated acetic acid solubilized into the oil phase. In
contrast, the RF samples with greater water content
contained more microbial loadings, which might be the
growth of acid tolerant microorganisms such as lactic acid
bacteria (Karas, Skvarča, & Žlender, 2002; Pourkomailian,
2000). Because of the activity of lactic acid bacteria, the pH
value of mayonnaise during storage deceased. In addition, it
has been reported that b-glucan concentrate used as fat
replacer in low-fat white-brined cheese resulted in a
significant lower pH value of the products than those of
their FF counterpart after 60 and 90 days storage (Volikakis
et al., 2004). It was assumed that possible fermentation of
the b-glucan concentrate must have occurred, resulting in
enhanced production of lactic acid. In this study, the pH
values of the RF mayonnaises after 64 days storage were
significantly (P%0.05) decreased with increasing the
b-glucan substitution.
3.3. Physical and textural characteristics
Table 3 shows the particle size of droplets, texture and
colour attributes of the FF and RF mayonnaises. The oil
droplet size of FF (9.09 mm) was the highest as compared
with that of the RF samples. Among the RF mayonnaises,
the 25B exhibited the largest particle size (3.62 mm) and the
50B gave the smallest droplet size (1.33 mm). This was a
result of different degree of shearing force applied during
the preparation processes. The FF and 25B contained such a
high oil content close to the critical volume fractions
(0.6–0.7 for monodisperse spheres and z0.8 for poly-
disperse emulsions) at which the spheres become close
packed (McClements & Demetriades, 1998). Within this
region, the emulsions are susceptible to undergoing a
composite phenomenon, called phase inversion, possibly
involving aspects of flocculation, coalescence, partial
coalescence, and emulsion formation in high shear fields
(McClements, 1999). Therefore, high shearing force, e.g.
high-pressure during homogenization cannot apply for the
FF and 25B formulations and hence the reduction of oil
droplet size was limited. For the 50B and 75B, more
intensive homogenization conditions were applied which
led to their smaller droplet size. Firmness and adhesiveness
of all samples showed similar trends. The values for FF and
75B were similar and slightly lower than those of the 50B,
whereas the 25B formulation exhibited only 50% the values
of 50B sample.
The lightness (L-value) of mayonnaise has a major
impact on the perceived appearance of the product. The L-
value of RF mayonnaises were lower than that of the FF
sample, except for the 50B formulation. Chantrapornchai,
Clydesdale, and McClements (1999) and McClements and
Demetriades (1998) reported that the emulsion changed
from a grey colour to an increasingly bright white colour as
the droplet size decreased due to an increase in light
scattering. Therefore, the L-value of 50B was significantly
higher than that of FF sample due to its much smaller
droplet size. The colour of b-glucan preparation (L*Z
43.63, a*Z3.91, and b*Z9.96) also influenced the colour
of RF mayonnaises. The a-values (redness) increased,
whereas the b-values (yellowness) decreased with increas-
ing the levels of b-glucan substitution. However, the
emulsion became less coloured as the droplet size decreased
(Chantrapornchai et al., 1999), this appeared to lower

Table 3
Physical properties of mayonnaise samples

Samples Particle size (mm) Texture Colour

Firmness (N) Adhesiveness L* a* b*
(!10K3 N m)
FF 9.09G0.16a 1.02G0.03b 23.82G0.84ab 82.31G0.59b K3.17G0.04d 23.34G0.58a
25B 3.62G0.05b 0.59G0.01c 12.50G0.15c 80.77G0.41c K0.80G0.01c 17.41G0.44b
50B 1.33G0.02d 1.12G0.02a 25.43G0.29a 83.92G0.06a K0.27G0.01b 12.37G0.04c
75B 1.93G0.11c 1.03G0.04b 22.51G0.85b 81.59G0.33bc 0.44G0.04a 11.67G0.11c

Assays were performed in triplicate. MeanGSD values followed by the same letter in each column are not significant different at P%0.05 by ANOVA and
Tukey’s test.
 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78
the effects of colour of b-glucan in the RF mayonnaises,
especially the 50B and 75B samples. In general, FF
mayonnaises exhibited a shiny bright yellow colour,
whereas RF mayonnaises were evaluated as too pale and
dense, mainly due to fat replacers, which is confirmed by the
previous finding (Karas et al., 2002).
3.4. Microstructure
The microstructures of FF and RF mayonnaise samples
are shown in Fig. 1. Close packing structures of spherical oil
droplets were observed in the FF and 25B samples. These
emulsions were easily distinguished by the presence of large
and uniform droplets (monodisperse) in the FF (Fig. 1(a)),
whereas the droplets of different size (polydisperse) in which
the small droplets were trapped between the large ones
occurred in the 25B sample (Fig. 1(b)). Gutierrez, Silva,
Chirinos, Leiva, and Rivas (2002) found that the viscosity of
polydisperse emulsions was significantly lower than that
observed in equivalent monodisperse emulsions at the same
volume fraction. Moreover, reduction of the fat content also
dramatically decreased viscosity of concentrated close
packing emulsions (McClements & Demetriades, 1998).
These can describe why the firmness and adhesiveness of the
25B was lower than those of the FF sample (Table 3). The
50B (Fig. 1(c)) and 75B (Fig. 1(d)) showed relatively open,
loose structures, with a network of aggregated droplets
containing a large number of interspaced voids of varying
Fig. 1. SEM micrographs of mayonnaise made with (a) FF, full fat; (b) 25B, 25% b-glucan substitution; (c) 50B, 50% b-glucan substitution; and (d) 75B, 75%
b-glucan substitution. Magnification is 2000!.
73
dimensions. These void spaces represented the aqueous
continuous phase in emulsions. The particulate b-glucans
and oil droplets seem to form a three-dimensional particle
network that extends throughout the volume of the emulsion.
The 75B containing lower oil content exhibited relatively
larger spaces surrounded by looser aggregated droplets than
that of the 50B sample. The smaller the droplet size, the
greater the extent of a three-dimensional gel network with the
more open structure formed at the lower oil volume fractions
leading to the larger emulsion viscosity (McClements, 1999).
This might also be attributed to the higher textural parameters
of 50B and 75B as compared with those of the 25B sample
(Table 3). In general, visual inspections of droplet sizes in
Fig. 1 are in good agreement with the data derived from
instrumental analysis using a particle size analyzer (Table 3).
3.5. Flow behaviour
Flow curves of the FF and RF mayonnaises are presented
in Fig. 2. All mayonnaise samples exhibited thixotropic
behaviour over the whole range of shear rate studied
(0–300 sK1) in that segments of the down curves reached
values of shear stress lower than those of the up curves at the
same values of shear rate. The thixotropic characteristics of
mayonnaise in the shear rate range of 0–200 sK1 were
reported earlier (Štern, Valentová, & Pokornỳ, 2001). The
power law model turned out to be best fitted to the
experimental shear rate–shear stress data (r2z0.99), expect
74 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78

350 10000

300

250
Shear stress (Pa)

1000
200

G', G'' (Pa)

150

100 100

50

0
0 50 100 150 200
Shear rate (s–1)
Fig. 2. Flow curves of mayonnaise samples; FF (&), 25B (:), 50B (C),
and 75B (%) measured at 25 8C. Full symbols represent the ascending part
and empty symbols the descending part of the flow curves.
for the up curve of the 75B sample of which the power law
parameters cannot be determined (Table 4). The consistency
coefficient (K) values were markedly decreased, whereas the
flow behaviour index (n) increased when the samples were
subjected to the ascending and then descending of the shear
rates. Therefore, the FF and RF mayonnaises behaved as
shear thinning fluids and their flow behaviour index (n)
values were less than 1.0. These findings are in agreement
with the results of Ma and Barbosa-Cánovas (1995).
In conclusion, all investigated mayonnaise samples
showed thixotropic shear thinning behaviour in that their
flow properties depended on both shear rate and time. In
concentrated emulsions, the droplets are close enough
together to interact with each other which may lead to the
formation of a three-dimensional network of aggregated
droplets. As the shear rate is increased, the hydrodynamic
forces cause aggregates to become deformed and eventually
disrupted which results in a reduction in the viscosity
(McClements, 1999).
250 300 10
0.1 1 10 100
Angular frequency (Hz)
Fig. 3. Dynamic mechanical spectra of mayonnaise samples; FF (&), 25B
(:), 50B (C), and 75B (%) measured at 0.5% strain and 25 8C. Full
symbols represent storage modulus (G 0 ) and empty symbols represent loss
modulus (G 0 0 ).
the separation of the two moduli (tan dZG 00 /G 0 ) is larger
than 0.1 (Fig. 4) (Clark & Ross-Murphy, 1987). The
previous studies also found that mayonnaise exhibited weak
gel characteristics in the frequency range of 0.1–10 Hz
(0.63–62.83 rad sK1) (Mancini et al., 2002). In general, it is
expected that emulsions with a greater fat content show
higher values of G 0 (Ma & Barbosa-Cánovas, 1995). In this
study, however, this effect was counterbalanced by the
substitution of b-glucan for fat in the formulation that
strengthens the gel structure of mayonnaises. Both 50B and
75B samples containing smaller droplets had higher G 0
value than the 25B and FF samples containing larger
droplets. These findings are in good agreement to those of
Langton, Jordansson, Altskär, Sørensen, and Hermansson
(1999) who reported that the linear viscoelastic properties as
measured by the G 0 values was found to be more solid-like
1

3.6. Dynamic rheological behaviour

Mechanical spectra of the FF and RF mayonnaises

obtained by the frequency sweep measurement classified
into that of weak gels because G 0 was larger than G 00
throughout the tested frequency range and both G 0 and G 00
tan δ

hardly correlated with the frequency change (Fig. 3) and

Table 4
Values of power law model parameters

Samples Up curves Down curves

2
K n r K n r2
0.1
FF 81.38 0.23 0.996 34.41 0.38 0.999
0.1 1 10 100
25B 26.05 0.33 1.000 15.38 0.42 0.991
50B 76.67 0.16 0.993 20.60 0.39 0.989 Angular frequency (Hz)
75B – – – 10.99 0.44 0.989
Fig. 4. Loss tangents (tan d) of mayonnaise samples; FF (&), 25B (:), 50B
K, consistency coefficient (Pa sn); n, flow behaviour index (dimensionless). (C), and 75B (%) measured at 0.5% strain and 25 8C.
 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78 75

when the mayonnaise was formed of smaller droplets (a) 30

produced by a higher energy input during processing, and
25
more liquid-like when it was formed of larger droplets
produced at lower energy input. Loss tangent (tan d)

Particle size (µm)

20
indicates whether elastic or viscous properties predominate
in a sample. The tan d values of all mayonnaise samples 15
(Fig. 4) were correlated with their oil droplet sizes (Table 3).
10
Samples with small droplets had lower tan d value and
exhibited more solids like behaviour than those with large 5
droplets. This result is in the same pattern with those of the
texture attributes (Table 3) except for the FF that exhibited (b) 0
relatively high texture attributes as compared with the 50B 1.2
and 75B samples in spite of the highest tan d value and
largest droplet size, probably because the close packing of 1

Firmness (N)
oil droplets that structure the material. 0.8
The steady-state flow curves obtained for these kind of
0.6
food emulsions (Fig. 2) are the result of dramatic structural
destruction, which may be attributed to both an irreversible 0.4
process (coalescence) and a reversible one (deflocculation).
0.2
A small amplitude oscillatory experiment, carried within the
linear viscoelastic region, on the other hand, has the (c) 0
advantage of minimizing destruction in the sample since
25
Adhesiveness (×10–3 Nm)

little or no permanent structure breakdown occurs during the

dynamic measurements. This approach allows a relationship 20
between the results obtained and the actual structure of
material to be drawn. Thus, the relationship between 15

complex viscosity and steady shear viscosity of the FF

10
and RF mayonnaises was established (Fig. 5). The steady
shear (large deformation) viscosity (h) of all mayonnaise 5
samples was apparently lower than the complex viscosity
(h*) (small deformation). This suggests that mayonnaise 0
0 10 20 30 40 50 60 70
behaviour was not described by the Cox–Merz rule, and this
Time (days)
phenomenon is typical of weak gels (Richardson, Morris,
Ross-Murphy, Taylor, & Dea, 1989). Fig. 6. Changes in oil droplet size (a) and instrumental texture attributes;
firmness (b) and adhesiveness (c) of mayonnaise samples; FF (&), 25B
(:), 50B (C), and 75B (%) during storage at room temperature. Symbols
100000 are means of three replicates and error bars represent the respective
standard deviations.

10000
3.7. Storage stability
1000
Mayonnaise stability was determined by monitoring
η*,η (Pas)

100 droplet sizes, texture attributes (Fig. 6), and microbiological

assays (Fig. 7) during storage at room temperature for
10 2 months. Storage did not produce significant effects on the
droplet size for all RF mayonnaises (Fig. 6(a)). However,
1 the FF mayonnaise showed a significant increase in droplet
size, especially after 40 days of storage at which the oil
0 separation was observed. This result confirms our previous
0 1 10 100 1000 finding that spent brewer’s yeast b-glucan can efficiently
Angular frequency, Shear rate (s–1) stabilize the oil-in-water emulsion (Thammakiti et al.,
2004). Kontogiorgos et al. (2004) reported that model salad
Fig. 5. Comparison between complex viscosity (h*) and steady shear
viscosity (h) of FF (&), 25B (:), 50B (C), and 75B (%) mayonnaise dressing samples creamed more than the ones containing
samples measured at 25 8C. Full symbols represent complex viscosity and cereal b-glucans; the latter creamed after 30 days of storage.
empty symbols represent steady shear viscosity. In a highly concentrated close packing emulsion like the FF
76
600
500
Total plate count (cfu/g)
400
300
200
100
0 3.8. Sensory evaluations
0 10 20 30
Time (days)
Fig. 7. Changes in total plate count of mayonnaise samples; FF (&), 25B
(:), 50B (C), and 75B (%) during storage at room temperature. Symbols
are means of three replicates and error bars represent the respective
standard deviations.
sample, an increase in droplet size possibly due to oil
droplet coalescence which occurred after the droplets had
been in contact for a prolonged period (Abu-Jdayil, 2003;
McClements, 1999). On the contrary, the stability of RF
mayonnaises was attributed to an increase in viscosity of the
continuous phase and a formation of weak gel network by
the added b-glucan (Thammakiti et al., 2004). Thus, the oil
droplets were kept apart by the particulate b-glucans and
coalescence became less likely than that of the FF sample.
In addition, it has been concluded that stability of a RF
mayonnaise can be improved by increasing energy input
during processing leading to the small droplets (Depree &
Savage, 2001; Franco, Berjano, Guerrero, Muňoz, &
Gallegos, 1995). Therefore, the good stability of RF
mayonnaises was attributed to the initial small droplets
which were stabilized by the added b-glucan preparation.
The firmness (Fig. 6(b)) and adhesiveness (Fig. 6(c)) of
all mayonnaise samples decreased rapidly during the first 10
days and after that remained nearly constant, except for the
FF sample which still decreased continuously. A similar
trend for mayonnaise-like emulsions was also observed by
Paraskevopoulou, Kiosseoglou, Alevisopoulos, and Kasapis
(1999). These authors reported a dramatic drop in emulsion
stability expressed in term of normalized G 0 values in the
first 24 h and followed by a much smoother consistency
reduction phase. This could be due to the very rapid
flocculation and/or coalescence of small droplets occurred
during the initial stage. The instability of texture attributes
Table 5
Sensory evaluation of mayonnaise samples
Samples Appearance
FF 7.27G1.23a
25B 5.50G1.22b
50B 4.60G1.48c
75B 4.00G1.39c
MeanGSD values followed by the same letter in each column are not significant different at P%0.05 by ANOVA and Tukey’s test. There were 30 panel lists in
the sensory panel.
S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78
for the FF sample, as expected, was due to the coalescence
of oil droplets which accelerated the creaming process. The
microbiological profiles of FF and RF mayonnaises during
storage were, as expected, presented in Fig. 7 and detail
discussion was reported in Section 3.2. In general, the total
plate counts of all mayonnaise samples prepared in this
study were in the acceptable limit according to the Thai
industrial standard (1000 cfu/g).
40 50 60 70
Sensory evaluation scores of the FF and RF mayon-
naises are shown in Table 5. The appearance and colour
scores significantly (P%0.05) decreased with increasing
the substitution levels of b-glucan. Colour of the RF
mayonnaises was evaluated as too pale and dense which is
confirmed by the data obtained from the instrumental
analysis (Table 3). The odour attribute, on the other hand,
was not significantly influenced by the b-glucan substi-
tution. The 25B and 50B samples showed no statistical
difference in texture and taste scores from the FF one,
whereas the 75B sample gave lower scores for these
attributes. The texture attribute was hardly correlated with
the data derived from instrumental analysis using the
texture analyzer (Table 3) and the rheometer (Figs. 3
and 4). The disparity between the highly variable and
usually high-speed compression between the teeth and the
constant slow compression speed of machines probably
accounts for some of the low correlations between sensory
and objective test, especially for strain rate sensitive foods
(Bourne, 2002). Štern et al. (2001) also found no
significant correlation between the apparent viscosity and
sensory characteristics of mayonnaise. The overall accept-
ability scores were significantly lower when the substi-
tution levels of b-glucan increased. This is mainly
contributed by the performance of appearance and colour
attributes. In this study, the overall acceptability scores of
three commercial mayonnaises evaluated by the same
sensory panel were ranged from 4.65 to 6.70 (data not
shown). Therefore, it is reasonable that the sensory
attributes with scores higher than 5 are considered
acceptable when compared with the controls. Thus, the
RF mayonnaises with a substitution level of b-glucan up to
50% were judged to be sensorially acceptable.
Colour Odour Texture Taste Acceptability
7.50G1.25a 5.53G1.55a 6.47G1.28a 6.18G1.55a 6.63G1.13a
5.13G1.20b 5.63G1.19a 5.80G1.19ab 5.40G1.45a 5.37G1.38b
4.37G1.50bc 5.07G1.23a 5.90G1.18ab 5.37G1.73a 5.03G1.40bc
3.67G1.32c 5.10G1.09a 5.53G1.38b 4.37G1.54b 4.45G1.46c
 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78
4. Conclusions
From the results of the present work, it can be concluded
that spent brewer’s yeast b-glucan played a multiple role as
a fat replacer and as well as an emulsion stabilizer which
resulted in the RF mayonnaises with lower calorie but
higher storage stability than their FF counterpart. However,
the addition of the b-glucan preparation adversely affected
mayonnaise appearance and colour leading to the signifi-
cantly lower sensory quality as compared with the FF
control sample. Nevertheless, the substitution levels of not
more than 50% of oil used were found to be acceptable.
Acknowledgements
This work was partially supported by a research grant
from Mahidol University and a scholarship from the Higher
Education Development Project—Agricultural Biotechnol-
ogy Consortium. The authors also thank Metrohm Siam Ltd,
Bangkok, Thailand for providing the Physica rheometer
used in these experiments.
References
Abu-Jdayil, B. (2003). Modelling the time-dependent rheological
behavior of semisolid foodstuffs. Journal of Food Engineering, 57,
97–102.
Akoh, C. C. (1998). Fat replacers. Food Technology, 52, 47–53.
AOAC (1990). (15th ed.) Official methods of analysis of the association of
official analytical chemists, Vol. II. Arlington, VA: Association of
Official Analytical Chemists.
Bligh, E. G., & Dyer, W. J. (1959). A rapid method of total lipid extraction
and purification. Canadian Journal of Biochemistry and Physiology, 37,
911–917.
Bourne, M. C. (1978). Texture profile analysis. Food Technology, 32, 62–
66 see also page 72.
Bourne, M. C. (2002). Food texture and viscosity: Concept and
measurement. New York: Academic Press.
Chantrapornchai, W., Clydesdale, F., & McClements, D. J. (1999).
Influence of droplet characteristics on the optical properties of colored
oil-in-water emulsions. Colloids and Surfaces A: Physicochemical and
Engineering Aspects, 155, 373–382.
Chirife, J., Vigo, M. S., Gómez, R. G., & Favetto, G. J. (1989). Water
activity and chemical composition of mayonnaises. Journal of Food
Science, 54, 1658–1659.
Clark, A. H., & Ross-Murphy, S. B. (1987). Structural and mechanical
properties of biopolymer gels. Advances in Polymer Science, 83, 57–
192.
Codex Alimentarius Commission (1989). Codex standard for mayonnaise
(Regional European Standard) CODEX STAN 168-1989. Codex
Coordinating Committee for Europe.
Depree, J. A., & Savage, G. P. (2001). Physical and flavour stability
of mayonnaise. Trends in Food Science and Technology, 12, 157–
163.
Egelandsdal, B., Langsrud, Ø., Nyvold, T., Sontum, P. K., Sørensen, C.,
Enersen, G., et al. (2001). Estimating significant causes of variation in
emulsions’ droplet size distributions obtained by the electrical sensing
zone and laser low angle light scattering techniques. Food Hydro-
colloids, 15, 521–532.
77
Franco, J. M., Berjano, M., Guerrero, A., Muňoz, J., & Gallegos, C. (1995).
Flow behaviour and stability of light mayonnaise containing a mixture
of egg yolk and sucrose stearate as emulsifiers. Food Hydrocolloids, 9,
111–121.
Gutierrez, X., Silva, F., Chirinos, M., Leiva, J., & Rivas, H. (2002).
Bitumen-in-water emulsion: An overview on formation, stability and
rheological properties. Journal of Dispersion Science and Technology,
23, 405–418.
Hathcox, A. K., Beuchat, L. R., & Doyle, M. P. (1995). Death of
enterohemorrhagic Escherichia coli O157:H7 in real mayonnaise and
reduced-calorie mayonnaise dressing as influenced by initial population
and storage temperature. Applied and Environmental Microbiology, 61,
4172–4177.
Karas, R., Skvarča, M., & Žlender, B. (2002). Sensory quality of standard
and light mayonnaise during storage. Food Technology and Biotech-
nology, 40, 119–127.
Klis, F. M., Mol, P., Hellingwerf, K., & Brul, S. (2002). Dynamics of cell
wall structure in Saccharomyces cerevisiae. FEMS Microbiology
Reviews, 26, 239–256.
Kogan, G. (2000). 1/3,1/6-b-D-Glucans of yeasts and fungi and their
biological activity. Studies in Natural Products Chemistry, 23, 107–
152.
Kontogiorgos, V., Biliaderis, C. G., Kiosseoglou, V., & Doxastakis, G.
(2004). Stability and rheology of egg-yolk-stabilized concentrated
emulsions containing cereal b-glucans of varying molecular size. Food
Hydrocolloids, 18, 987–998.
Langton, M., Jordansson, E., Altskär, A., Sørensen, C., & Hermansson, A.
(1999). Microstructure and image analysis of mayonnaises. Food
Hydrocolloids, 13, 113–125.
Ma, L., & Barbosa-Cánovas, G. V. (1995). Rheological characterization of
mayonnaise. Part II: Flow and viscoelastic properties at different oil and
xanthan gum concentrations. Journal of Food Engineering, 25, 409–
425.
Mancini, F., Montanari, L., Peressini, D., & Fantozzi, P. (2002). Influence
of alginate concentration and molecular weight on functional properties
of mayonnaise. Lebensmittel-Wissenschaft Und-Technologie, 35, 517–
525.
McClements, D. J. (1999). Food emulsions; principles, practice, and
techniques. Boca Raton, FL: CRC Press.
McClements, D. J., & Demetriades, K. (1998). An integrated approach to
the development of reduced-fat food emulsions. Critical Reviews in
Food Science and Nutrition, 38, 511–536.
Mohamed, M. O., & Morris, H. A. (1987). Textural and microstructural
properties of rennet-induced milk coagulum as affected by the
addition of soy protein isolate. Journal of Texture Studies, 18, 137–
155.
Paraskevopoulou, A., Kiosseoglou, V., Alevisopoulos, S., & Kasapis, S.
(1999). Influence of reduced-cholesterol yolk on the viscoelastic
behaviour of concentrated o/w emulsions. Colloids and Surfaces B:
Biointerfaces, 12, 107–111.
Pourkomailian, B. (2000). Sauces and dressings. In D. Kilcast, & P.
Subramaniam (Eds.), The stability and shelf-life of food. Washington,
DC: CRC Press.
Radford, S. A., & Board, R. G. (1993). Review: Fate of pathogens in home-
made mayonnaise and related products. Food Microbiology, 10, 269–
278.
Reed, G., & Nagodawithana, T. W. (1991). Yeast-derived products.2nd ed.
Yeast technology (pp. 369–440). New York: Van Nostrand Reinhold.
Richardson, R. K., Morris, E. R., Ross-Murphy, S. B., Taylor, L. J., & Dea,
I. C. M. (1989). Characterization of the perceived texture of thickened
systems by dynamic viscosity measurements. Food Hydrocolloids, 3,
175–191.
Roland, I., Piel, G., Delattre, L., & Evrard, B. (2003). Systematic
characterization of oil-in-water emulsions for formulation design.
International Journal of Pharmaceutics, 263, 85–94.
Souci, S. W., Fachman, W., & Kraut, H. (2000). Food composition and
nutrition tables (6th ed.). Stuttgart: Medpharm.
78 S. Worrasinchai et al. / Food Hydrocolloids 20 (2006) 68–78

Štern, P., Valentová, H., & Pokornỳ, J. (2001). Rheological properties and
sensory texture of mayonnaise. European Journal of Lipid Science and
Technology, 103, 23–28.
Stone, B. A., & Clarke, A. E. (1992). Chemistry and biology of (1/3)-
b-glucans (pp. 525–564). Victoria, Australia: La Trobe University
Press.
Thammakiti, S., Suphantharika, M., Phaesuwan, T., & Verduyn, C. (2004).
Preparation of spent brewer’s yeast b-glucans for potential applications
in the food industry. International Journal of Food Science and
Technology, 39, 21–29.
Volikakis, P., Biliaderis, C. G., Vamvakas, C., & Zerfiridis, G. K. (2004).
Effects of a commercial oat-b-glucan concentrate on the chemical,
physico-chemical and sensory attributes of a low-fat white-brined
cheese product. Food Research International, 37, 83–94.
Xiong, R., Xie, G., & Edmondson, A. S. (2000). Modelling the pH of
mayonnaise by the ratio of egg to vinegar. Food Control, 11, 49–56.