Food Research International xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Food Research International

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

Foaming properties and foam structure of milk during storage

⁎
Thao M. Ho, Thi Huynh Anh Le, Anna Yan, Bhesh R. Bhandari, Nidhi Bansal
ARC Dairy Innovation Hub, School of Agriculture and Food Sciences, The University of Queensland, Australia

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

Keywords: Changes in foaming properties and foam structure of raw whole, raw skim, pasteurised whole, and pasteurised
Foamability skim milk during storage at 4 °C, evaluated by mechanical mixing, air and steam injection, were investigated.
Foam stability The results showed that storage of milk until the end of their shelf-life (day 3 for raw milk and day 21 for
Foam structure pasteurised milk) did not induce any significant change in pH, particle size, viscosity and their foaming prop-
Milk storage
erties (foaming capacity, foam stability and size of air bubbles) although there was a slight increase in free fatty
Foaming method
Shelf-life
acid content. Regarding foaming methods, the air injection method produced much less foam volume than
mechanical mixing and steam injection, except for raw milk where mechanical mixing showed the least foam-
ability. However, an opposite trend was observed for foam stability. Although air injection induced the largest
size of air bubbles, it produced the most stable foam, which was much more stable than foam produced by
mechanical mixing and steam injection.

1. Introduction
Foam can be defined as a two-phase colloidal system in which the
gas bubbles are dispersed in the continuous liquid phase (Walstra
1989). In the dairy industry, foam properties of milk are of a major
concern for many products, typically cappuccino-style drinks. For these
products, characteristics of a foam layer on the top determine the
quality of the whole product and the attraction of consumers because of
its essential contribution to appearance, volume, texture, mouthfeel and
release rate of coffee aroma (Khezri, Shahriari, & Shahsavani 2017).
However, control of foaming process to create the foam with desirable
characteristics is extremely complicated and still is a challenge to re-
searchers and dairy industry although many studies on the foaming of
milk has been conducted since the beginning of 20th century. This is
because of the inherent instability of the foam which leads to difficul-
ties in the determination of foam properties, and high dependence of
foam behaviour on numerous factors, ranging from the source, quality
and composition of milk to milk processing and storage conditions
(Huppertz 2010). There have been a substantial number of studies on
effects of milk composition (e.g. lipid content, free fatty acids, protein
concentration), processing conditions (e.g. pH, homogenisation, heat
treatment prior to foaming, foaming temperature), foaming methods
and additives on foaming properties of cow milk (Borcherding,
Chrlorenzen, & Hoffmann 2009; Borcherding, Hoffmann, Lorenzen, &
Schrader 2008; Borcherding, Lorenzen, Hoffmann, & Schrader 2008;
Gamboa & Barraquio 2012; Goh, Kravchuk, & Deeth 2009; Kamath
2007; Kamath, Huppertz, Houlihan, & Deeth 2008; Kamath,
Wulandewi, & Deeth 2008; Khezri et al. 2017; Marinova et al. 2009;
Martinez-Padilla, Garcia-Mena, Casas-Alencaster, & Sosa-Herrera 2014;
Oetjen, Bilke-Krause, Madani, & Willers 2014; Pijnenburg, Sala,
Valenberg, & Meinders 2012; Ward, Goddard, Augustin, & McKinnon
1997; Zhang, Dalgleish, & Goff 2004), just to name a few. In the
meantime, to the best of our knowledge, only a few research works
have been dedicated to changes in foaming properties of milk during
storage. In this regard, Deeth and Smith (1983) reported that storage of
milk at 4 °C for 2 days did not significantly affected its foamability.
However, Gamboa and Barraquio (2012) found that aging of milk at
4–11 °C for 9 days resulted in an significant decrease in foam capacity,
but did not affect the foam stability. The differences in the reported
results could be due to the dissimilarities in foaming methods (steam
injection for the former and whipping for the latter) and aging time. It is
noticed that the milk used in these studies are pasteurised and homo-
genised milk, and the aging period is much shorter than the shelf-life of
milk. Moreover, these studies are limited to the evaluation of foam-
ability and foam stability. Therefore, a further investigation on changes
in foaming properties of different types of milk (e.g. raw and processed
milk) during storage over their shelf-life is required.
Theoretically, any activity which can introduce gas (air) into the
bulk of milk could can be utilized to create foam, such as pouring,
shaking, sparkling, bubbling, mixing, agitating, supersaturating,

⁎
Corresponding author at: School of Agriculture and Food Sciences, The University of Queensland, Brisbane, QLD 4072, Australia.
E-mail addresses: minh.ho1@uq.net.au (T.M. Ho), l.anh@uq.edu.au (T.H.A. Le), anna.yan@uq.net.au (A. Yan), b.bhandari@uq.edu.au (B.R. Bhandari),
n.bansal@uq.edu.au (N. Bansal).

https://doi.org/10.1016/j.foodres.2018.08.051
Received 19 June 2018; Received in revised form 8 August 2018; Accepted 18 August 2018
0963-9969/ © 2018 Elsevier Ltd. All rights reserved.

Please cite this article as: Ho, T.M., Food Research International (2018), https://doi.org/10.1016/j.foodres.2018.08.051
T.M. Ho et al.
whipping or beating (Walstra 1989; Wilde & Clark 1996). Typically, in
most coffee shops, the milk foam is produced via direct injection of
steam into the milk through a very small opening or nozzle placed just
under the milk surface. The most desirable characteristics of this
method are to heat the milk to 65-70 °C which is an indispensable re-
quirement for dispersing of hot beverages and to partially denature the
proteins which contributes to the enhancement of foam stability (Silva
et al. 2008). However, this foaming technique is very difficult to con-
trol, highly decided by practical experience of a steaming operator, and
easy to cause an overheating of the foaming milk which results in a loss
of milk flavour as well as harsh and bitter undertones to the foam
surface because of the extreme denaturation of proteins and scalding of
milk. Thus, by employing this foaming approach, each person can
create a foam with different properties. Moreover, added water into the
milk from condensed steam, which might account for 10–12% (v/v) of
total volume of collapsed foam and liquid under foam layer, can affect
foam stability (Deeth & Smith 1983). For the sake of research, there are
similar other foaming approaches which are easy to operate, control
and implement. They are air injection and mechanical mixing. In the
former, a certain volume and/or a fixed flow rate of air is forced into a
fixed volume of liquid milk through an aperture, a frit, an orifice, a
diffuser stone or a glass sinter with a desirable porosity. This foaming
method does not induce any change in the milk temperature (Wilde &
Clark 1996). Thus, in order to mimic the steam injection approach, the
milk is preheated to a desired temperature prior to subjecting to
foaming. In mechanical mixing, high speed movement of milk-whisking
devices, high-shear mixers or blenders provide the shear force to cut the
milk surface and to incorporate the air into the bulk of milk, by which
foam is produced (Huppertz 2010). With an advancement in tech-
nology, an integration of a heating unit into mechanical mixer so that
heating and foaming process can simultaneously happen, is commer-
cially available. Each foaming method is different in its operating me-
chanism, subsequently creating foam with different properties (Goh
et al. 2009; Silva et al. 2008). Therefore, it is necessary to include
several foaming methods to evaluate the foaming properties of milk.
In this study, changes in foaming properties of four types of liquid
milk (raw whole, raw skim, pasteurised whole, and pasteurised skim
milk) during storage at 4 °C over their shelf-life were investigated using
three common foaming methods (steam injection, air injection and
mechanical mixing).
2. Materials and methods
2.1. Materials
Raw whole milk, pasteurised whole milk and pasteurised skim milk
were provided by a local commercial milk processing factory
(Queensland, Australia). Pasteurised whole and skim milk were
homogenised at 140 bar for 1st stage and 30 bar for 2nd stage by the
processor. Raw skim milk was produced by separating cream from the
raw whole milk. The raw whole milk initially pre-heated in a water
bath to 35 °C was poured into a commercial cream separator (Motor
Sich 100-18 Separator, Motor Sich, Ukraine, capacity 80–100 L/h, drum
rotational speed 10,500 ± 100 RPM and normal working temperature
35–40 °C) for cream separation. The cream was discarded while the
skim milk was collected for further experiment. After the milk was re-
ceived, the cream separation was immediately carried out. All analy-
tical grade chemicals used in this study were purchased from Sigma
Aldrich (New South Wales, Australia).
2.2. Effect of storage on properties of milk and foam
Four milk samples, raw whole milk, raw skim milk and their re-
spective pasteurised milk, in plastic bottles (high-density polyethylene),
were straightway kept in cold room at 4 °C, after collection, without
light expose to evaluate effects of storage period on foaming properties.
2
Food Research International xxx (xxxx) xxx–xxx
Depending on the shelf-life of each types of milk, the time to take the
samples for analysing was different. For raw milk, the samples were
analysed at day 0 and 3 while for pasteurised milk, the experiments
were conducted over a 21-day storage on day 1, 7, 14 and 21 which was
the end of their shelf-life.
2.3. Determination of milk properties
2.3.1. Free fatty acid content analysis
The free fatty acid (FFA) content of all milk was determined using
the titration method described by Deeth and Fitz-Gerald (1975). A
mixture of isopropanol, n-hexane and 4 N sulphuric acid was used to
extract FFA in about 3 mL milk. The extracted solution was then titrated
with 0.02 N methanolic potassium hydroxide to determine the amount
of FFA (μ.equiv./mL).
2.3.2. Determination of particle size
Particle size of milk was measured using a laser light scattering
analyser (Malvern Mastersizer 2000, Malvern Instruments Ltd.,
Worcestershire, UK). The samples were slowly added to distilled water,
which was used as dispersant, until an obscuration rate of 15% was
reached in the particle size analyser. The particle size of the samples
was calculated based on Mie theory with refractive indices of milk fat
and water being 1.462 and 1.330 respectively, and with an assumption
of spherical particles. The volume weighted mean D[4,3], and surface
area weighed mean diameter, D[3,2] (μm) were used to interpret the
particle size data (Truong, Bansal, & Bhandari 2014).
2.3.3. pH measurement
The pH of milk samples was measured using a pH Cube Benchtop
(TPS Pyt Ltd., Australia) with range from 0 to 14 pH and accuracy +/−
0.01 pH at the specific time points. The meter was calibrated with
buffers at each time of measurement.
2.3.4. Viscosity
Viscosity of milk was measured with an AR 1500 Rheometer (TA
Instruments, UK), using a cone and plate geometry (cone diameter
40 mm, angle 0, gap 0.2 mm). For each measurement, 2.0 mL sample
was carefully deposited over the plateau of the rheometer. Steady state
flow measurements were carried out at 25 ± 0.1 °C in the range of
0–100 s−1. The rheological parameters (shear stress, shear rate and
apparent viscosity) were obtained from the software of TA.
2.4. Foaming methods
2.4.1. Steam injection
Foaming of milk by steam injection was as per reported by Kamath
(2007) with a slight modification using Café Series EM6910 (Sunbeam,
Australia), and experiences from barista training. The equipment was
turned on at least 30 min prior to experiment. About 100 mL of milk
was poured into a 250 mL graduated plastic jug (Genetics Australia Co-
operative Ltd., Australia) and then cooled down to 5 °C using ice bath.
Prior to performing the foaming process, the steam wand was flushed to
remove any condensed water. Initially, the milk jug was placed on a
mechanically adjustable stage and the steam wand was positioned at an
angle of 45o respective to the milk surface and the tip of steam wand
was located about 2 mm under the milk surface. The adjustable stage
had a round holder to firmly hold the milk jug, and a handle to move
the milk jug up and down steadily. During foaming process, in order to
maintain an appropriate contact point of the steam wand tip and milk
surface due to a rapid change in the boundary between foam and liquid,
the milk jug was continuously moved and the rate of the movement was
based on the sound of air incorporation. Once the steam was turned on,
the milk jug was moved down (vertically) gradually to maintain the
contact point of the steam wand and milk surface to keep introducing
air into the bulk of milk until the milk temperature reached to 45 °C.
T.M. Ho et al.
Then, the milk jug was immediately moved up (vertically) to lower the
steam wand tip into the milk to texturise and heat the milk up to 65 °C.
The milk temperature was continuously measured using a digital
thermal meter.
2.4.2. Mechanical mixing
A Breville cafe milk frother (BMF600, Australia) was used in this
study and foaming procedure was carried out by following the in-
structions in manufacturer's booklet with a slight modification (Breville
2017). This equipment is integrated with a heating unit allowing to
heat the milk while foaming. About 250 mL of the milk which was in-
itially cooled to 5 °C (using ice bath) was poured into the frother jug. A
“capp” frothing disc was used to produce the foam suitable for cap-
puccino-style drinks. Continuous rotation of the frothing disc helps to
incorporate the air so that the foam is produced. The foaming process
was automatically stopped when the milk temperature was about 65 °C
(approximately 2 min).
2.4.3. Air injection
An air injection foaming apparatus reported by Kamath, Huppertz,
et al. (2008) was slightly modified and used in this study. The system
comprised of a stainless steel diffusion air stone (0.5 μm in pore size,
2.5 cm in length, 1.27 cm in diameter, Ibrew, Australia) which was
connected to a plastic tube (high-density polyethylene, 0.5 m in length,
Pmax = 10 bar, 1.25 cm in diameter) at one end. The other end of a
plastic tube was attached to air pressurization system through a flow
meter (Dwyer, Model RMA-11-SSV, New South Wales, Australia, range
50–200 mL/min air) and a pressure regulator (SMC, USA, range
0–200 kPa). For foaming, 100 mL of milk was measured in a 250 mL
graduated plastic jug and pre-heated to 65 °C in water bath. The air
stone was entirely submerged in the milk and up to a distance of about
1 cm from the bottom of the jug. Air was injected at a pressure of
45–50 kPa and a flow rate of 130–135 mL/min for 25 s.
Prior to carrying out the experiments, the reproducibility and re-
peatability of all foaming methods were checked via testing the foam-
ability of reconstituted low heat skim milk powder (8.5 g/L) at the same
above-mentioned foaming conditions for each method.
2.5. Determination of foam properties
2.5.1. Foamability and foam stability
Foamability was expressed as foam ratio (−) which is a ratio of
foam volume (mL) and volume of milk sample used to foam (100 mL for
steam and air injection and 250 mL for mechanical mixing). For steam
and air injection, the foam volume was directly measured from the
graduated plastic jug. For mechanical mixing, the foam was poured into
a 500 mL plastic cylinder to determine the foam volume, and the foam
remaining in the frother jug was scraped as much as possible into the
measuring cylinder using a plastic spatula. The measurement was per-
formed immediately as the interfacial layer between liquid and foam
3
Food Research International xxx (xxxx) xxx–xxx
was observed.
Foam stability at room temperature (25 °C) was measured by the
time (min) taken for the foam to collapse. Typically, the cappuccino-
style drinks are consumed to half their volume within 10 min after
being processed, thus foam stability was evaluated via the measurement
of total volume of foam retained after 10 min. Foam stability was ex-
pressed as the percentage reduction in foam volume after 10 min (%)
which was calculated as per Eq. 1. In Eq. 1, VF0 and VF10 are volume of
foam at t = 0 and t = 10 min, respectively.
VF 0 − VF 10
%reduction in VF after 10 min = ∗100
VF 0 (1)
2.5.2. Foam structure
Image of the foam surface was taken using a light microscope (Prism
Optical, Eagle Farm, Queensland, Australia) which was fitted with a 5.0
MP camera system using TSView7 software (Fuzhou Tucsen Image
Technology Co., Ltd., China) connected to the video entry port of a
computer. An Olympus LG-PS2 lamp (Eagle Farm, Queensland,
Australia) was used to illuminate the foam. Images captured at 0 and
10 min of foaming were analysed in terms of diameter of air bubbles
using Image-Pro Plus 6.0 software (Media Cybernetics Inc., Bethesda,
USA). Because all air bubbles were not spherical, the longest length was
considered as the diameter of an air bubble.
2.6. Experimental design and statistical analysis
The experiments were performed following a completely rando-
mized design with three replications with three different milk batches
and two measurements were performed for each replicate. The ex-
perimental data were subjected to analysis of variance (ANOVA) at
significance level p = 0.05 using the Minitab 16.0 statistical pro-
gramme. For some criteria, due to similarity in analytical data of milk at
each time point, only analytical results at day 0 and the end of shelf-life
were shown. The statistical analysis was performed over the shelf-life
for each type of milk.
3. Results and discussion
3.1. Changes in milk properties during storage
3.1.1. FFA and pH
The changes in FFA and pH of raw whole and skim milk, and their
pasteurised counterparts during storage period are shown in Fig. 1. For
FFA (Fig. 1a), at day 0 the raw milk had a higher FFA content than its
pasteurised counterpart. The higher FFA value of the raw milk is be-
cause they do not undergo pasteurisation which inactivates milk lipase.
The inactivation of milk lipase prevents lipolysis from occurring and
consequently releasing FFA (Deeth 2006). A removal of a large pro-
portion of the fat during the processing of the raw and pasteurised skim
Fig. 1. Changes in FFA content (a) and pH (b) of the
milk over storage period. In these figures, ( )
and ( ): Raw skim milk; ( ) and
( ): Raw whole milk; ( ) and
( ): Pasteurised skim milk; and ( )
and ( ): Pasteurised whole milk.
T.M. Ho et al.
milk resulted in a lower FFA content as compared to FFA in raw and
pasteurised whole milk. During storage, FFA content increased in all
samples at every time point from day 0 to the end of shelf-life. How-
ever, for both raw skim and whole milk there was no significant dif-
ference in FFA content between day 0 and 3. For the pasteurised milk,
the differences in the FFA content between whole milk and skim milk
became more significant as the storage period was increased. The FFA
of the pasteurised whole milk increased at a faster rate than that of
pasteurised skim milk and reached to 1.67 μ.equiv./mL on day 21,
which are almost threefold more than the FFA in the pasteurised skim
milk (0.63 μ.equiv./mL) at the end of shelf-life. The negative effects of
FFA on foaming properties of milk have been reported (Buchanan 1965;
Deeth & Smith 1983; Kamath, Wulandewi, & Deeth 2008). Thus, the
increase of FFA during storage might cause poor foaming in milk.
As reported previously, the change in pH of raw and pasteurised
whole milk during storage at 4 °C were associated with production of
lactic acid caused by the growth of microorganisms (Bermúdez-Aguirre,
Mawson, Versteeg, & Barbosa-Cánovas 2009; Ravanis & Lewis 1995;
Ziyaina, Govindan, Rasco, Coffey, & Sablani 2018). As can be seen in
Fig. 1b, all milk types had a similar pH value (6.60–6.70) and there was
almost no change during storage. These results indicated that the
above-mentioned changes did not occur in the samples over their shelf-
life. Similarly, it was reported that when kept at 4 °C, raw and pas-
teurised milk did not show any change in pH up to day 5 and 16, re-
spectively (Bermúdez-Aguirre et al. 2009). An alteration of pH leads to
changes in the charge, structure and ionic strength of proteins (Ward
et al. 1997), or in the amount of caseins, calcium and phosphate pre-
senting in the serum (Huppertz 2010). All of these changes affect the
availability of surface active substances, the rheological properties of
interfacial films and the milk viscosity, and subsequently foamability of
milk.
3.1.2. Particle size
The changes in particle size in all milk during storage are illustrated
and Table 1 and in Fig. S1 (supplementary material). These results in-
dicated that the fat globules in the milk did not undergo any sig-
nificantly changes during the storage period. The differences in particle
size of fat globules among the milk samples are due to homogenisation
and cream separation during processing. The raw whole milk had the
largest particle size with D[4,3] and D[3,2] being 4.30 and 3.29 μm,
respectively while pasteurised skim milk had the smallest one with
0.14 μm for D[4,3] and 0.12 μm for D[3,2]. The particle size distribu-
tion of raw whole milk was dominated by that of milk fat globules while
the particle size distribution of skim milk represented mainly the casein
micelles. Particle size distribution of pasteurised whole milk in Fig. S1
was bi-modal as expected, the large peak was for homogenised fat
globules and the smaller peak was contributed by the casein micelles. It
can be seen in Fig. S1, the small peak in the particle size distribution of
pasteurised whole milk was almost identical to that of skim milk
Table 1
Particle size of fat globules in terms of D[4,3] and D[3,2] in all milk over storage period⁎.
Samples Day 0 Day 3
D[4,3]⁎⁎ D[3,2]⁎⁎⁎ D[4,3] D[3,2]
Raw skim milk 0.21 ± 0.02a 0.12 ± 0.00a 0.27 ± 0.09a 0.12 ± 0.01a
Raw whole milk 4.30 ± 0.04a 3.29 ± 0.09a 4.29 ± 0.07a 3.28 ± 0.10a
Pasteurised 0.14 ± 0.00a 0.12 ± 0.01a NA NA
skim milk
Pasteurised 0.54 ± 0.01a 0.23 ± 0.01a NA NA
whole milk
⁎
Means with different letters (e.g. a letters) in the same row indicated significant differences during storage (p < 0.05).
⁎⁎
D[4,3]: Volume weighed mean diameter.
⁎⁎⁎
D[3,2]: Surface area weighed mean diameter.
⁎⁎⁎⁎
NA: Not analysed.
Food Research International xxx (xxxx) xxx–xxx
samples.
For particle size diameter, it seems to be that D[3,2] was more
correlated to particle size distribution than D[4,3] because the particle
size distribution of raw and pasteurised skim milk was almost the same
while only D[3,2] exhibited similar values for raw and pasteurised skim
milk (Fig. S1). Due to larger size of fat globules, flocculation into large
aggregates was visually observed for the raw whole milk at day 3, but
floccules were easy to disperse back in the milk via slight mixing. Milk
fat globules in raw skim milk, and pasteurised whole and skim milk are
insufficient or too small to induce any creaming or coalescence, thus
particle size of these samples was almost constant during storage.
Reduction of particle size of milk fat via homogenisation helps to
improve foamability and foam stability because once homogenised, the
solid fat crystals in the fat globules probably become too small to dis-
rupt the foam lamella (Huppertz 2010). However, the effects of
homogenisation on foaming properties are dependent on temperature
at which foam was created. The results from previous studies by
Borcherding, Hoffmann, et al. (2008), Deeth and Smith (1983) and
Kamath, Huppertz, et al. (2008) suggest that foaming properties of milk
were markedly enhanced by homogenisation as foaming process was
done at 10-40 °C, but when foaming process was performed at tem-
peratures higher 50 °C, homogenisation only marginally affected the
foaming properties of milk. In this study, all three foaming methods
were carried out at 65 °C, it is expected that foamability and foam
stability of raw whole milk and pasteurised whole milk is similar de-
spite the differences in their fat globule size due to homogenisation.
3.1.3. Viscosity
Viscosity of all milk samples at day 0 and the end of shelf-life are
shown in Fig. 2. As expected, at day 0 the raw and pasteurised whole
milk had a higher viscosity than the corresponding skim milk. This
difference is likely caused by the differences in the fat content in milk.
The viscosity of milk increased with an increase in fat content (Li,
Joyner, Carter, & Drake 2018). However, at the end of shelf-life the
viscosity of raw skim and whole milk was almost constant whereas that
of pasteurised skim and whole milk significantly increased (Fig. 2). This
result was comparable to the findings reported by Ismail and El Deeb
(1973) in which it was stated that the viscosity of pasteurised milk
increased upon storing at 5 °C up to 48 h.
However, at low shear rate (e.g. 50 (1/s) in Table S1 (supplemen-
tary material) the viscosity of all milk at day 0 was around 2 mPa.s and
did not increase significantly during storage. The measured viscosity of
all milk was comparable to that reported by Brisson and Singh (2013) in
which viscosity of milk at 20 °C has been estimated to be around
1.79 mPa.s for skim milk and 2.13 mPa.s for whole milk. In foaming,
effects of viscosity on foamability and foam stability is opposite in
which an increase of viscosity results in a reduction in foamability, but
an enhancement in foam stability (Martínez-Padilla, García-Rivera,
Romero-Arreola, & Casas-Alencáster 2015). Nonetheless, the viscosity
Day 7 Day 14 Day 21
D[4,3] D[3,2] D[4,3] D[3,2] D[4,3] D[3,2]
NA⁎⁎⁎⁎ NA NA NA NA NA
NA NA NA NA NA NA
0.14 ± 0.00a 0.12 ± 0.00a 0.14 ± 0.00a 0.12 ± 0.00a 0.14 ± 0.00a 0.11 ± 0.01a
0.54 ± 0.01a 0.24 ± 0.01a 0.54 ± 0.02a 0.23 ± 0.01a 0.53 ± 0.01a 0.23 ± 0.00a
4
T.M. Ho et al.
of milk when foamed is greatly dependent on foaming temperature and
solids concentration (Borcherding et al. 2009; Kamath 2007).
3.2. Changes in foam properties and foam structure during storage
3.2.1. Foamability and foam stability
The foamability and foam stability, which were investigated by
three foaming approaches, of all milk over their shelf-life are shown in
Fig. 3. For foamability (Fig. 3a) at day 0, regardless of the foaming
5
Food Research International xxx (xxxx) xxx–xxx
Fig. 2. Viscosity of milk samples at day 0 (a) and the
end of shelf-life (b). In these figures, ( ): Raw
skim milk; ( ): Raw whole milk; (
): Pasteurised skim milk; and ( ):
Pasteurised whole milk.
method, the raw whole milk exhibited a lower foam capacity than the
other milk. This difference was much more apparent in mechanical
mixing foaming technique in which foam ratio of the raw whole milk
was almost fourfold less than that of the other milk. The lowest foam-
ability of the raw milk is likely because of high fat content, high FFA
concentration and large particle size of fat globules (Fig. 1 and Table 1).
It is noticed that both pasteurised whole and skim milk were homo-
genised by which the size of fat globules reduced significantly. An as-
sociation of a poor foaming of milk with fat and FFA content, and
Fig. 3. Changes in foaming ability (a) and foam
stability (b) of all milk samples during storage in
three foaming methods; (1) - mechanical mixing, (2)
- air aeration, and (3) - steam injection. In each
figure, means with different letters in the same type
of milk indicated significant differences during sto-
rage (p < 0.05); ( ): Raw skim milk;
( ): Raw whole milk; ( ):
Pasteurised skim milk; and ( ): Pasteurised
whole milk.
T.M. Ho et al.
particle size of fat globules has been well reported (Borcherding,
Hoffmann, et al. 2008; Gamboa & Barraquio 2012; Kamath,
Wulandewi, & Deeth 2008). While fat and FFA molecules interfere with
the formation of intermolecular interactions of proteins, resulting in
weakening or destroying the integrity, cohesiveness and viscoelastic
properties of the interfacial films; the large fat crystals in the large fat
globules can puncture the interfacial films (Wilde & Clark 1996). In-
terestingly, during storage over the shelf-life, although there was a
fluctuation in the foamability of all milk, these changes were not sig-
nificant (p > 0.05). This result is compatible to that reported by Deeth
and Smith (1983) in which it was reported that the foaming capacity of
pasteurised milk did not change during aging at 4 °C for 48 h, and by
Nanua, Osorio, Mcgregor, and Traylor (2004) in which it was observed
that the foam value of pasteurised whole milk showed no change during
storage at 3 °C for 15 days. However, Gamboa and Barraquio (2012)
reported that after a 9-day storage period of pasteurised skim milk,
there was a decrease in the amount of foam produced.
For foam stability (Fig. 3b), the higher percentage of foam volume
reduction after 10 min (%) expresses the lower foam stability. Similar to
foamability, the foam produced from the raw whole milk had the least
stability regardless of foaming method, as compared to that created
from the other milk. During storage, percentage of foam volume re-
duction after 10 min of all milk did not increase significantly
(p > 0.05). It was reported that changes in foaming properties of milk
during aging at refrigerated temperature were caused by hydrolysis of
milk proteins (as a result of enzymes produced by microorganisms)
which led to a decrease in pH, and by lipolysis of milk fat which re-
sulted in an increase in FFA content (Celestino, Iyer, & Roginski 1996;
Gamboa & Barraquio 2012).
The reasons for insignificant changes in foaming properties of all
milk during storage could be that, as presented in above section, during
storage the physical properties of milk in terms of pH, viscosity and
particle size of fat globules which govern the foaming properties of milk
hardly altered or changed while the increase in FFA of pasteurised milk,
especially pasteurised whole milk, possibly is not sufficient to induce
any change in foaming properties of milk. Kamath, Wulandewi, and
Deeth (2008) reported that the milk still exhibited a good foamability
and a high foam stability at FFA concentration < 2 μ.equiv./mL while
the maximum amount of FFA at the end of shelf-life was about
1.6 μ.equiv./mL (pasteurised whole milk).
As can be seen in Fig. 3b, the percentage of foam volume reduction
after 10 min in three foaming methods, especially for steam injection
was relatively high (approximately 25–50% depending on foaming
method and type of milk), indicating that a large proportion of foam
volume collapsed within 10 min after foam was created. The high
percentage of foam volume reduction after 10 min in this study could be
because the measurements to evaluate foamability and foam stability
were performed immediately after foaming as soon as the interfacial
layer was observed (about 15 s after foaming). It was reported by
Jimenez-Junca, Sher, Gumy, and Niranjan (2015) that the drainage of
liquid from the foam produced by steam injection happened rapidly
within initial few minutes after foaming. After 2–4 min, > 80% (v/v) of
the amount of liquid incorporated into foam during foaming process
were drained. The higher rate of drainage liquid from the foam is re-
sponsible for the lower foam stability.
3.2.2. Foam structure
Bubble size distribution and representative microscopic images of
the surface of foam produced from the raw whole and skim milk, and
their pasteurised counterparts under steam injection, immediately after
foam formation (t = 0) and after 10 min (t = 10) are illustrated in
Figs. 4. At day 0 and for all foams, visual observation revealed that at
t = 0 foams were polyhedral-shape with uniform size and well-defined
lamellae. However, after 10 min (t = 10), the number of bubbles de-
creased and many big air bubbles were observed. This change is a result
of coalescence and disproportionation of air bubbles which are initiated
6
Food Research International xxx (xxxx) xxx–xxx
by the drainage of liquid from the foam (Walstra 1989). These foam
destabilization processes were easily observed in images of foam at
t = 10 min. Moreover, the mean diameter values of foams at t = 10 min
were higher than those at t = 0, and the size distribution curve of air
bubbles also shifted from the left to the right in all samples (Fig. 4).
Similar changes in characteristics of the foam at t = 0 and t = 10 were
also observed for the milk at the end of shelf-life. An increase in the size
of air bubbles during storage of foam (from t = 0 to t = half-life) was
also reported (Kamath, Huppertz, et al. 2008). Similar properties of
bubble size distribution in foam produced by air aeration and me-
chanical mixing foaming methods are illustrated in Figs. S2 and S3
(Supplementary material).
For all foaming methods and most types of milk, the aging of milk
until their shelf-life resulted in an insignificant increase in the size of air
bubbles. The increase in FFA content during storage could be a reason
for increase in size of air bubbles. It was found by Kamath, Wulandewi,
and Deeth (2008) that coarseness of foams increased with increasing
FFA content of milk. Among three foaming methods, the size of air
bubbles produced by mechanical mixing was smallest while air injec-
tion produced the largest air bubbles. However, the size of air bubbles is
highly dependent on many parameters of foaming process (such as
porosity of air stone, air pressure and flow rate in air injection; type of
nozzle, steam pressure in steam injection; and type of frother disc in
mechanical mixing).
3.3. A comparison of foaming methods in terms of foamability and foam
stability
A comparison on foamability and foam stability of three foaming
methods for different types of milk is illustrated in Fig. 5. For foam-
ability, because a different amount of milk was used in different
foaming methods, a foam ratio which is a ratio of initial foam volume
(mL) to amount of milk (mL) was calculated to compare the foamability
among them. As shown in Fig. 5(a), the foamability of different milk
was very dissimilar for different foaming methods. For the raw whole
milk, while the mechanical mixing showed a very poor foamability
(foam ratios ≈0.3–0.6), both the air and steam injection approaches
gave very high foam ratio values (≈1.0–1.5). For raw skim milk, pas-
teurised whole and skim milk, although air injection showed the least
foaming capacity, as compared to both steam injection and mechanical
mixing which had similar foam ratios, the differences in foam ratio
among three foaming methods were not significant. These results in-
dicated that mechanical mixing is unsuitable to produce foam from the
raw whole milk which has a high fat content and large particle size of
fat globules. As reported, below 40 °C, milk fat exists as a mixture of
liquid fat and solid fat crystals with sharp edges and corners. Under
mechanical forces, the solid fat crystals can pierce the membrane of fat
globules to release liquid fat over the interface. Liquid fat can adsorb
rapidly on the interfacial regions, but unable to form high viscoelastic
films to stabilize air bubbles. Therefore, at below 40 °C, the negative
effects of milk fat on the foam properties become more predominantly
evident (Jimenez-Junca et al. 2015; Walstra 1989). In mechanical
mixing method, it took about 2 min to heat the milk to from 5 to 65 °C
during foaming while the required time for steam injection was only
30 s and the milk was preheated to 65 °C in air injection foaming
method. A slow heating of milk in the mechanical mixing, as compared
to that in steam injection, likely makes the effect of milk fat on foaming
capacity more pronounced as the melting rate of milk fat was slower.
For foam stability (Fig. 5b), for all types of milk, the foam produced
from air injection was much more stable than that created by steam
injection and mechanical mixing. A lower stability of foam produced by
mechanical mixing and steam injection than air injection is possibly a
result of pouring the foam into a measuring cylinder and the unsteady
movement of the container during foaming, respectively, by which the
foam structure can be disturbed. Moreover, it seems that heating the
milk prior to subjecting to foaming enhances foam stability. The pre-
T.M. Ho et al.
heating of milk up to 65 °C not only induces partial denaturation of
proteins by which foam stability is significantly enhanced (Silva et al.
2008), but also liquefies all solid fat crystals in the milk fat by which the
negative effect of milk fat on foam stability is minimized (Pijnenburg
et al. 2012). In a similar study reported by Goh et al. (2009), it was
stated that among three foaming methods (steam injection, air bubbling
and mechanical agitation) the stability of whole milk foam was similar
while skim milk foam generated by mechanical agitation was sig-
nificantly more stable than that created by the other two approaches.
The differences in results from this study and those reported previously
are because of the differences in foaming conditions in each method as
well as dissimilarities in types of used milk. In a foaming approach, for
example in steam injection, the changes in process parameters such as
7
Food Research International xxx (xxxx) xxx–xxx
Fig. 4. Images (a and b), size distribution of air
bubbles (c and d) and table illustrating average size
of air bubbles of foam produced by steam injection at
day 0 and the end of shelf-life of milk. Scale
bar = 1000 μm; (–) and (–): Raw skim milk at t = 0
and 10 min; ( ) and ( ): Raw whole milk at t = 0
and 10 min; ( ) and ( ): Pasteurised whole milk
at t = 0 and 10 min; and ( ) and ( ): Pasteurised
skim milk at t = 0 and 10 min.
steam pressure and nozzle design, leads to differences in foamability
and foam stability and even size of air bubbles and foam texture
(Jimenez-Junca et al. 2015).
4. Conclusion
The study showed that aging of milk at 4 °C over its shelf-life (day 3
for raw whole and skim milk, and day 21 for pasteurised whole and
skim milk) did not induce any significant change in pH and particle size
of fat globules. Despite slight increase in FFA content and viscosity
during aging, the foaming properties in terms of foamability, foam
stability, appearance of foam, and size of air bubbles were almost un-
changed during storage. Moreover, different foaming approaches
T.M. Ho et al.
Fig. 5. A comparison of foaming ability and stability of three foaming methods.
In these figures, ( ): Raw skim milk; ( ): Raw whole milk; ( ): Pasteurised
skim milk; and ( ): Pasteurised whole milk.
produced the foam with different properties. While mechanical mixing
approach was unsuitable to produce foam from the raw whole milk and
created the smallest air bubbles, the most stable foam was found in air
injection foaming technique.
Acknowledgement
This research was funded through Australian Research Council‘s
Industrial Transformation Research Hub (ARC-ITRH) grant. The ARC
Dairy Innovation Hub is a collaboration between The University of
Melbourne, The University of Queensland and Dairy Innovation
Australia Ltd. The authors acknowledge the facilities, and the scientific
and technical assistance, of the School of Agriculture and Food Sciences
at The University of Queensland.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.foodres.2018.08.051.
References
Bermúdez-Aguirre, D. A., Mawson, R., Versteeg, K., & Barbosa-Cánovas, G. V. (2009).
Composition properties, physicochemical characteristics and shelf life of whole milk
after thermal and thermosonication treatments. Journal of Food Quality, 32, 283–302.
Borcherding, K., Chrlorenzen, P., & Hoffmann, W. (2009). Effect of protein content,
Food Research International xxx (xxxx) xxx–xxx
casein-whey protein ratio and pH value on the foaming properties of skimmed milk.
International Journal of Dairy Technology, 62, 161–169.
Borcherding, K., Hoffmann, W., Lorenzen, P. C., & Schrader, K. (2008). Effect of milk
homogenisation and foaming temperature on properties and microstructure of foams
from pasteurised whole milk. LWT- Food Science and Technology, 41, 2036–2043.
Borcherding, K., Lorenzen, P. C., Hoffmann, W., & Schrader, K. (2008). Effect of foaming
temperature and varying time/temperature-conditions of pre-heating on the foaming
properties of skimmed milk. International Dairy Journal, 18, 349–358.
Breville (2017). Breville the milk café instrution booklet, Breville, Australia. 9–13. https://
breville.com.au/collections/milk-frothers/products/the-milk-cafe?variant=
35157711954.
Brisson, G., & Singh, H. (2013). Milk composition, physical and processing characteristics.
In R. Chandan, & A. Kilara (Eds.). Manufacturing yogurt and fermented milks (pp. 21–
48). Oxford: John Wiley & Sons.
Buchanan, R. (1965). Lipolysis and the frothing of milk. Australian Journal of Dairy
Technology, 20, 62–66.
Celestino, E., Iyer, M., & Roginski, H. (1996). The effects of refrigerated storage on the
quality of raw milk. Australian Journal of Dairy Technology, 51, 59–63.
Deeth, H., & Fitz-Gerald, C. (1975). A convenient method for determining the extent of
lipolysis in milk. Australian Journal of Dairy Technology, 30, 109–111.
Deeth, H. C. (2006). Lipoprotein lipase and lipolysis in milk. International Dairy Journal,
16, 555–562.
Deeth, H. C., & Smith, R. A. D. (1983). Lipolysis and other factors affecting the steam
frothing capacity of milk. Australian Journal of Dairy Technology, 38, 14–19.
Gamboa, G. V., & Barraquio, V. L. (2012). Foaming properties at different fat levels and
age of milk. Philippine Agricultural Scientist, 95, 416–421.
Goh, J., Kravchuk, O., & Deeth, H. C. (2009). Comparison of mechanical agitation, steam
injection and air bubbling for foaming milk of different types. Milchwissenschaft-Milk
Science International, 64, 121–124.
Huppertz, T. (2010). Foaming properties of milk: A review of the influence of composition
and processing. International Journal of Dairy Technology, 63, 477–488.
Ismail, A. A., & El Deeb, S. A. (1973). Effect of heat processing, storing and homo-
genization on the viscosity, opacity and stability of cow and buffalo milks. Zeitschrift
für Lebensmittel-Untersuchung und -Forschung, 152, 202–207.
Jimenez-Junca, C., Sher, A., Gumy, J.-C., & Niranjan, K. (2015). Production of milk foams
by steam injection: The effects of steam pressure and nozzle design. Journal of Food
Engineering, 166, 247–254.
Kamath, S. (2007). Foaming of milk, university of Queensland. University of Queensland.
Kamath, S., Huppertz, T., Houlihan, A. V., & Deeth, H. C. (2008). The influence of tem-
perature on the foaming of milk. International Dairy Journal, 18, 994–1002.
Kamath, S., Wulandewi, A., & Deeth, H. (2008). Relationship between surface tension,
free fatty acid concentration and foaming properties of milk. Food Research
International, 41, 623–629.
Khezri, M., Shahriari, S., & Shahsavani, L. (2017). The effect of xanthan gum and tem-
perature on foam stability of milk-based espresso coffees. Journal of Food Biosciences
and Technology, 7, 15–22.
Li, Y., Joyner, H. S., Carter, B. G., & Drake, M. A. (2018). Effects of fat content, pas-
teurization method, homogenization pressure, and storage time on the mechanical
and sensory properties of bovine milk. Journal of Dairy Science, 101(4), 2941–2955.
Marinova, K. G., Basheva, E. S., Nenova, B., Temelska, M., Mirarefi, A. Y., Campbell, B., &
Ivanov, I. B. (2009). Physico-chemical factors controlling the foamability and foam
stability of milk proteins: Sodium caseinate and whey protein concentrates. Food
Hydrocolloids, 23, 1864–1876.
Martínez-Padilla, L., García-Rivera, J., Romero-Arreola, V., & Casas-Alencáster, N.
(2015). Effects of xanthan gum rheology on the foaming properties of whey protein
concentrate. Journal of Food Engineering, 156, 22–30.
Martinez-Padilla, L. P., Garcia-Mena, V., Casas-Alencaster, N. B., & Sosa-Herrera, M. G.
(2014). Foaming properties of skim milk powder fortified with milk proteins.
International Dairy Journal, 36, 21–28.
Nanua, J., Osorio, L., Mcgregor, U., & Traylor, S. (2004). Contribution of the products of
fat hydrolysis to the frothing capacity of milk. International Journal of Agriculture and
Rural Development, 5, 114–120.
Oetjen, K., Bilke-Krause, C., Madani, M., & Willers, T. (2014). Temperature effect on
foamability, foam stability, and foam structure of milk. Colloids and Surfaces a-
Physicochemical and Engineering Aspects, 460, 280–285.
Pijnenburg, J., Sala, G., van Valenberg, H., & Meinders, M. B. (2012). Foaming behaviour of
organic and regular milk. Research report - No. 429945. (Wageningen UR).
Ravanis, S., & Lewis, M. J. (1995). Observations on the effect of raw milk quality on the
keeping quality of pasteurized milk. Letters in Applied Microbiology, 20(3), 164–167.
Silva, S., Espiga, A., Niranjan, K., Livings, S., Gumy, J. C., & Sher, A. (2008). Formation
and stability of milk foams. In G. M. Campbell, M. G. Scanlon, & D. L. Pyle (Eds.).
Bubbles in Food 2: Novelty, Health and Luxury (pp. 153–162). Minnesota: AACC
International, Inc.
Truong, T., Bansal, N., & Bhandari, B. (2014). Effect of Emulsion Droplet size on Foaming
Properties of milk Fat Emulsions. Food and Bioprocess Technology, 7, 3416–3428.
Walstra, P. (1989). Principles of foam formation and stability. In A. Wilson (Ed.). Foams:
Physics, chemistry and structure (pp. 1–15). London: Springer.
Ward, B. R., Goddard, S. J., Augustin, M. A., & McKinnon, I. R. (1997). EDTA-induced
dissociation of casein micelles and its effect on foaming properties of milk. Journal of
Dairy Research, 64, 495–504.
Wilde, P. J., & Clark, D. C. (1996). Foam formation and stability. In G. M. Hall (Ed.).
Methods of testing protein functionality (pp. 110–152). London: Blackie Academic and
Professional.
Zhang, Z., Dalgleish, D. G., & Goff, H. D. (2004). Effect of pH and ionic strength on
competitive protein adsorption to air/water interfaces in aqueous foams made with
mixed milk proteins. Colloids and Surfaces. B, Biointerfaces, 34, 113–121.
Ziyaina, M., Govindan, B. N., Rasco, B., Coffey, T., & Sablani, S. S. (2018). Monitoring
shelf life of pasteurized whole milk under refrigerated storage conditions: Predictive
models for quality loss. Journal of Food Science, 83(2), 409–418.