Innovative Food Science and Emerging Technologies 34 (2016) 6876

Contents lists available at ScienceDirect

Innovative Food Science and Emerging Technologies

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

Impact of cryoconcentration on casein micelle size distribution, micelles

inter-distance, and ow behavior of skim milk during refrigerated storage
Alseny Balde a,b, Mohammed Aider a,c,
a
b
c

Institute of Nutrition and Functional Foods (INAF), Universit Laval, Quebec G1V0A6, Canada
Department of Food Sciences, Universit Laval, Quebec G1V0A6, Canada
Department of Soil Sciences and Agri-Food Engineering, Universit Laval, Quebec G1V0A6, Canada

a r t i c l e

i n f o

Article history:
Received 28 October 2015
Received in revised form 30 December 2015
Accepted 31 December 2015
Available online 27 January 2016
Keywords:
Cryoconcentration
Skim milk
Casein micelles
Size
Rheology
Color

a b s t r a c t
Cryoconcentration combined with a cascade effect was used to concentrate skim milk up to 25.12% total dry matter. Size, shape, and inter-micellar distance of casein micelles were characterized by ZetasizerNano-ZS, transmission electron microscopy, and ImageJ analyses. Flow properties of the cryoconcentrated skim milk were
evaluated during 5 weeks of storage under refrigerated condition at 4 C. Milk color was also evaluated according
to the L*, a*, and b* system. The cryoconcentrated skim milk obtained after three cryoconcentration cycles was
characterized by a monomodal distribution of its micelles with a tendency to smaller casein micelles. Approximately 60% of the total micellar volume was occupied by the casein micelles with a size of 100200 nm, less
than 18% of the volume with a size of 50100 nm and only less than 1% was occupied by micelles with a size
N 350 nm. This result shows that cryoconcentration changed the distribution of the mean size of the casein micelles to smaller units. No signicant difference was observed on the inter-micellar distance. Cryoconcentration
signicantly improved the color of skim milk by increasing the L* value up to 67 which was similar to that of
whole milk. Transition from a Newtonian to a non-Newtonian behavior was observed from the fourth week storage with a slight increase of casein micelle size.
Industrial relevance: A concentration procedure of skim milk based on a complete block cryoconcentration technique was proposed. Application of this sub-zero technology permitted the concentration of skim milk total dry
matter up to 25%. The casein micelle size was positively affected by moving the major part of the micelles toward
the smaller size, whereas the inter-micellar distance was not affected. This new knowledge can be exploited in
milk-based products to enhance the product stability. The cryoconcentrated skim milk color was positively
affected since its L* value, which represents the milk whiteness, was signicantly improved. The ow behavior
of the cryoconcentrated milk was of Newtonian type up to 4 weeks of storage at 4 C. The generated knowledge
in this study can be easily used by the milk processing industry in order to make stable milk product with high
dry matter content without adding milk powder, which negatively affects the product sensory properties (oury
consistency).
2016 Elsevier Ltd. All rights reserved.

1. Introduction
The use of concentration processes in dairy industry can signicantly
contribute to enhance the overall efciency of milk processing since
huge quantities of milk can be reduced by concentrating the total dry
matter of some specic components such as proteins, yielding advantages in terms of processing, packaging, transportation, and handling
Keshani, Luqman Chuah, Nourouzi, Russly, & Jamilah, 2010. The selection of a convenient concentration process depends on the required
level of concentration, the impact of the process on products quality,
Corresponding author at: Department of Food Sciences, Universit Laval, Quebec
G1V0A6, Canada. Tel.: +1 418 656 2131 #4051; fax: +1 418 656 3723.
E-mail address: mohammed.aider@fsaa.ulaval.ca (M. Aider).

http://dx.doi.org/10.1016/j.ifset.2015.12.032
1466-8564/ 2016 Elsevier Ltd. All rights reserved.

available energy resources, and the relative cost of the process. Sometimes, a combination of different concentration processes is also used
(Morison & Hartel, 2006). Currently, there are several concentration
methods available for enhancing milk concentration such as vacuum
evaporation, reverse osmosis, ultraltration, and cryoconcentration in
its different variants (freeze concentration) Miyawaki, Liu, Shirai,
Sakashita, & Kagitani, 2005. Cryoconcentration of skim milk is a process
of concentrating the solid matter contained in a the aqueous phase
by removing part of water in a form of ice Aider, de Halleux, &
Melnikova, 2009. The ice formation can be achieved by different ways
such as suspension crystallization, progressive freeze concentration,
and complete block cryoconcentration Gunathilake, Dozen, Shimmura,
& Miyawaki, 2014; Iritani, Katagiri, Okada, Cao, & Kawasaki, 2013.
Among these techniques, the complete block cryoconcentration is the

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

simplest one for milk concentration and it is based on a controlled freezing followed by a controlled passive or assisted thawing.
This technology, being operated at sub-zero temperature conditions,
is attractive for heat sensitive liquid foods since it allows retaining the
nutritional quality and aromatic compounds within the product. This
particularity is attributed to the low operating temperatures, which
are suitable to avoid the degradation of the sensitive liquid food components such as heat labile proteins, vitamins, and volatiles (Flesland,
1995; Ghizzoni, Del Popolo, & Porretta, 1995, and the absence of a liquidvapor interface. Actually, it is regarded as highly promising separation process of water from liquid foods without affecting the quality and
properties of other components (Fellows, 2000). Considering these
advantages, cryoconcentration technology has been investigated by
several researchers for its application to a variety of liquid foods, such
as milk, milk whey, fruit juices, maple sap, and saline solutions Di
Cesare, Cortesi, & Martini, 1993. Cryoconcentrated milk is a dairy product that may be utilized as an intermediate material for sterilized,
sweetened condensed milk, and Greek-type yoghurt production or as
a nal product for the consumer or use in different food formulations.
However, even if cryoconcentration technology seems to offer several advantages in comparison with other concentration techniques such
as heat evaporation and membrane concentration (Thijssen, 1970;
Thijssen & Van Der Malen, 1981), a high dehydration of casein micelles
by water removal processes is of particular importance (De Kruif, 1999;
Morris, Foster, & Harding, 2000; Walstra, 1979), since it can increase the
volume fraction of dispersed particles and the inter-micelles interactions Bienvenue, Jimenez-Flores, & Singh, 2003. The latter (inter-micelles interactions) is very important since the casein micelles have
the greatest impact on the milk macroscopic and functional properties
Liu, Dunstan, & Martin, 2012, and they are the main contributors to
the viscosity of milk (Walstra & Jenness, 1984) and signicantly inuence the cheese yield. Thus, any factor that alters the aggregation state
of casein micelle, such as pH, concentration, and salt balance, undoubtedly affect the viscosity of milk (Bienvenue et al., 2003). In skim milk,
the continuous phase viscosity is largely determined by lactose concentration, whereas the volume fraction of suspended material is determined by proteins such as casein micelles, dissociated caseins, native
whey proteins, and denatured whey proteins (Anema, 2008; Jeurnink
& De Kruif, 1993). It has been reported that when cryoconcentration is
used, the increase of the total dry matter in the concentrated phase is accompanied by an increase of the amount of lactose entrapped in the ice
crystals (Aider & Ounis, 2012). However, little is currently known how
the physico-chemical properties of the casein micelles change in response to cryoconcentration.
A better understanding of the physico-chemical properties of
cryoconcentrated milk and the changes occurring with progressively
increasing cryoconcentration level is needed. This is necessary to
further understand the dynamics of structure changes during
cryoconcentration and ultimately better determine the main principles ruling the processing of cryoconcentrated milk under different
storage conditions. Many studies have been conducted on the viscosity of concentrated milk prepared by heat evaporation (Vlez-Ruiz &
Barbosa-Cnovas, 1998), ultraltration Karlsson, Ipsen, Schrader, &
Ard, 2005, or powder reconstitution Alexander, Rojas-Ochoa, Leser,
& Schurtenberger, 2002; Anema, 2008; Dahbi, Alexander, Trappe,
Dhont, & Schurtenberger, 2010, and some ow properties of freezeconcentrated skim milk were reported (Chang & Hartel, 1997). By contrast, limited information is available on the effects of freezing procedures such as the cryoconcentration on the micelle size, the intermicellar distance (spacing between casein micelles), and the product
ow properties during storage.
Hence, the purpose of this study is (1) to evaluate how
cryoconcentration combined with a cascade affect inuences the
size of casein micelles, the inter-micelle distance, as well as the
cryoconcentrated skim milk color, and (2) to establish the impact
of the cryoconcentration procedure on the rheological properties of

69

cryoconcentrated skim milk during refrigerated storage. The evaluation

of the inuence of temperature and concentration on the apparent viscosity was also performed during 5 weeks of storage at 4 C.
2. Materials and methods
2.1. Skim milk and cryoconcentration procedure
Pasteurized skim milk was purchased from Natrel (Agropur Cooperative, Quebec, Canada) and was used as the start material. Skim milk
proximate composition was the following: total dry matter, 9.24
0.15%; lactose, 4.91 0.21%; total protein, 3.54 0.17%; and ash content, 0.79 0.11%. The initial pH value was 6.5 0.15.
The cryoconcentration procedure was carried out by applying the
cascade principle (effect) as reported by Aider and Ounis (2012). To
do this, glass bottles of 9.3 cm inner diameter and 15 cm height were
used. The bottles were lled with 1000 ml skim milk and were frozen
in a freezer at 20 2 C. After 24 h of freezing, the thawing step
was carried out under simple gravitational separation of the concentrate
from the ice block at ambient temperature (23 1 C). The collected
fraction was maintained at near-zero temperature by immersing the
collection bottle in ice water. This procedure avoided any risk of bacterial growth during the thawing period. At this step, 500 mL of the
initial frozen volume was thawed and collected (50% of the initial
volume). This fraction constituted the concentrated phase of the
1st cryoconcentration cycle. The same procedure was repeated with
the collected concentrated solution, which was used as a feed solution for the 2nd cryoconcentration cycle. At the end of the second
cryoconcentration cycle, 50% of the thawed solution was collected.
This procedure was repeated at the 3rd cryoconcentration cycles.
Each concentrate at a given cycle was used as feed solution for the
next cycle. To avoid any bacterial growth during the storage of the collected samples, sodium azide (0.01% w/w) was added to the skim milk
as a preservative.
2.2. Analyses
2.2.1. Total dry matter
Total dry matter for all samples was determined by measuring
weight loss upon drying in an oven at 100 C under vacuum until constant weight (24 h) and expressed as dry matter content/total weight
in %. The accuracy of the measurements was veried by a freeze drying
control of some samples. No noticeable difference was nd between the
measurements.
2.2.2. Total proteins
Total protein content of each sample was determined by Dumas
combustion method by using an FP-528 Leco apparatus (Leco Corporation, St. Joseph, MI, USA). The instrument was calibrated with
ethylenediaminetetraacetic acid (EDTA) as a nitrogen standard. The
percentage of total protein content was calculated from nitrogen
content by multiplying by a factor 6.38 (IDF, 2002).
2.2.3. Ash and mineral fraction analysis
Ash content of different skim milk samples and ice fractions was
measured by the incineration method in a mufe furnace at 550 C for
20 h. The specic mineral analysis (Ca, P, Mg, Na, and K) was carried
out by the inductively coupled plasma method (ICP, Optima 4300 DV,
Perkin-Elmer, Norwalk, CT, USA) with the following wavelengths:
317.933, 396.847, and 393.366 nm for Ca; 285.213, 280.271, and
279.553 nm for Mg; 766.490 nm for K; and 589.592 and 588.995 nm
for Na (Carnovale, Britten, Couillard, & Bazinet, 2015).
2.2.4. Concentration factor and process efciency
The concentration factor for each component (proteins, ash, specic
minerals) at each cryoconcentration cycle was calculated as a ratio of its

70

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

content in the cryoconcentrated fraction at a given cycle (Cn) to its content in the initial feed skim milk solution (C0) by the following equation
(Eq.(1)).
Cf

Cn
 100%
C0

where Cf is the concentration factor for specic component (%), Cn is the

corresponding concentration at a given cryoconcentration cycle (n = 1,
2, and 3), and C0 is the specic component concentration in the initial
skim milk solution.
Cryoconcentration process efciency was calculated by the following equation (Eq.(2)):
DMn
PE
 100%
DMn1

where DM(n) is the total dry matter content at a given cryoconcentration

cycle, and
DM(n-1) is the total dry matter corresponding to the preceding
cryoconcentration cycle.
Eq. (2) was adapted to calculate the process efciency for each component by replacing the DM(n) and DM(n-1) by the concentration of each
targeted component (proteins, ash, and specic minerals).
The recovery efciency of total dry matter, proteins, and specic
minerals was calculated as follows (Eq. (3))
Reff

C i  C i1
100%
C i1

where Ci is the concentration of the targeted component at the ith

cryoconcentration cycle (i = 1, 2, 3), and C(i 1) is its concentration in
the previous cycle.
2.2.5. Color measurement
The color parameter measurements of the initial and
cryoconcentrated skim milk were carried out according to the
method reported by Hwang et al. (2007), after calibrating its original
values with a standard plate (Y = 92.6, x = 0.3162, y = 0.3323). L*,
a*, and b* values were regarded as indicator of lightness, redness, and
yellowness, respectively. A volume of about 12.5 ml of milk was placed
in a glass cuvette, inserted into a black chamber (provided by Konica
Minolta), and connected to the chromameter. The color of each milk
sample was determined at room temperature using a chromameter
(Konica Minolta CR-300, Morinuchi, Tokyo, Japan) and recorded in the
CIE-Lab tristimulus system. Each measurement was taken in triplicate
for each sample and average values were used.
2.3. Transmission electronic microscopy
After a cryoconcentration procedure was completed, a small drop
(50 L) of each sample was immediately xed by adding 2.5% glutaraldehyde in cacodylate buffer (0.1 M, pH 7.3) for 24 h at 4 C. The samples

were then dehydrated with serial concentrations of ethanol (30%, 50%,

70%, 95%, and 100%). The samples were embedded in Epon, sectioned
with a diamond knife in an ultra-microtome (6080 nm), stained with
uranyl acetate (3%) and lead citrate (0.1%), and observed by using a
transmission electron microscopy (JEOL JEM-1230 JEOL Ltd., Tokyo,
Japan) at 80 kV with a magnication factor of 12,000 with Gatan Microscopy Suite (Digital Micrograph software, Version 2.11.1404.0). Untreated milk was observed with the same manner and used as a control
sample.
2.4. Image analysis
For each sample, at least 20 transmission electron micrographs were
taken. The distance between the casein micelles was measured by using
ImageJ software (ImageJ, Version 1.48v). The mean values of the distance measured at ve different sides were taken as representative
values of each sample. All calculated values were statistically analyzed
and validated.
2.5. Particle size determination
The average particle size of casein micelles was measured by using
the Malvern ZetasizerNano-ZS instrument (Malvern Instruments Ltd.,
Worcestershire, UK). The instrument is equipped by a laser with a
wavelength of 633 nm. The intensity of scattered light is detected at
the 173 angle. Cryoconcentrated milk from different cycles was diluted
200 fold and ltered (0.45 m nylon lters). Milk permeate was analyzed. A volume of 1500 L of each sample was inserted into a singleuse disposable sizing cuvette DTS0012 following the method reported
in the work of Mootse et al. (2014). The light-scattering measurements
were conducted at 25 C and initiated 4 min after dilution. All measurements were performed in triplicate.
2.6. Rheological measurements
The rheological measurements were conducted with a Physica
Rheometer (Physica-Rheolab ARES-G2, TA Instruments, Wood Dale,
IL, USA) by using concentric cylinder (DIN) geometry. Experimental
studies were conducted on skim milk. Skim milk with 3.54% protein
and 90.75% moisture was used to make concentrates of 14.76%,
21.36%, and 25.12% total dry matter for rheometric measurements
at 5 C, 15 C, and 25 C during 5 weeks of storage at 4 C. The three
selected temperatures represent refrigeration, intermediate, and atmosphere conditions, respectively, as possible temperatures in
which the concentrated milk may be stored and handled by industries or by the consumers in home conditions. An aliquot volume of
12.5 mL sample of cryoconcentrated milk was loaded in the cup of
Rheometer. The sample was pre-sheared at 500 s 1 for 60 s, and
shear rate was monitored over a range of 101000 s 1 during
12 min.

Table 1
Proximate composition of initial and cryoconcentrated skim milk and process efciency as a function of the cryoconcentration cycle.
Component

Total solids (w/w)

Total Ca, mg/100 g
Total P, mg/100 g
Total K, mg/100 g
Total Na, mg/100 g
Total Mg, mg 100 g
Protein, % (w/w)
Ash, % (w/w)

Initial skim milk

9.24 0.01
107.3 1.6
84.7 2.2
97.2 1.7
54.1 1
9.1 0.2
3.54 0.17
0.72 0.01

Concentration cycle

Process efciency, %

1st

2nd

3rd

1st

2nd

3rd

14.73 0.03
175.9 2.8
139.4 3.8
167.1 3.0
87.4 1.7
15.2 0.4
5.66 0.07
1.4 0.01

21.36 0.04
249.1 3.9
197.2 4.7
250.1 6.0
119.6 2.9
22.4 0.7
7.73 0.24
1.52 0.02

25.12 0.12
325.9 5.0
249.9 6.9
332.3 6.3
137.4 2.6
28.6 0.4
9.02 0.13
1.94 0.12

159.42 31
163.93 27
164.58 18
171.91 15
161.55 17
167.03 20
159.89 7
194.44 9

231.17 35
232.15 31
232.84 19
257.30 13
221.07 13
246.15 27
218.36 12
211.11 13

271.86 42
303.73 15
295.04 23
341.87 11
253.97 19
314.29 12
254.80 18
269.44 11

Table 2
Color parameters of the cryoconcentrated skim milk.
Color
parameter

Initial skim milk

L*
a*
b*

62.68 0.25a
4.98 0.04c
3.525 0.02d

Concentration cycle
1st

2nd

3rd

66.32 0.75b
6.1 0.12c
0.44 0.13e

66.96 0.16b
6.63 0.04c
1.58 0.08e

66.97 0.36b
7.37 0.07c
2.67 0.08e

Data are presented as mean standard deviation.

Both simple power law and HerschelBulkley models were used to

describe the shear rateshear stress data for high-concentration skim
milk:
:n

Powerlaw : K

:n

HerschelBulkley models : 0 K

Efficiency of the main minerals recovery, %

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

71

120

100

80

72

67

64

65

61

C0

60
50

C1

47

42

40

C2

41
36

32

30

C3

28

26

20

15

0
Ca

Mg

Na

Type of the targeted minerals

where is the yield stress, K is the consistency index (N sn/m2), and n is

the ow behavior index.

Fig. 2. Effect of the cryoconcentration on the recovery of different milk minerals calculated
by the equation (Eq. (3)).

2.7. Statistical analysis

These results are in agreement with those reported by Aider et al.

(2009). The highest process efciencies were recorded at the end of
the rst and second cryoconcentration cycles. This result indicated
that the recovery of the total dry matter was the highest at these cycles.
However, the results obtained in the present study showed smaller dry
matter recovery compared to the results reported in the study of Aider
and Ounis (2012), who reached higher concentration levels (Aider &
Ounis, 2012). This quantitative differences could be attributed to some
handling procedures and the geometry of the used devices. The initial
total protein content of the skim was 3.54% 0.17%. By increasing the
cryoconcentration cycle, the total protein content increased in a quasilinear mode and reached average values of 5.66% 0.07%, 7.73%
0.24%, and 9.02% 0.12% and the end of the 1st, 2nd, and 3rd
cryoconcentration cycle, respectively. These values represent an increase of protein content of 68.4, 145, and 233% at the rst, second,
and third cryoconcentration cycle, respectively, compared with the initial total protein content.

All experiments had three replications for each treatment and measurement. Data were presented as means and standard deviations (SD).
Analysis of variance (ANOVA) was performed, and the mean comparisons were carried out by Tukey's test at 95% condence level. Statistical
analysis was performed using the SAS 9.3 software (SAS Institute Inc.,
Cary, NC, USA).
3. Results and discussion
3.1. Dry matter, total protein content, and concentration factor
The chemical composition of the initial skim milk and
cryoconcentrated milk fractions is summarized in Table 1. The initial total dry matter content of the skim milk was 9.24% 0.01%.
By increasing the cryoconcentration cycle, the total dry matter content increased approximately linearly and reached 14.73% 0.03%,
21.36% 0.04%, and 25.12% 0.12% at the end of the rst, second,
and third cryoconcentration cycle, respectively. The evolution of the
total dry matter can be described by a linear regression equation as
follows:
TDM 5:4243  CrS 4:0539

with R2 = 0.9895 and where TDM is the total dry matter (%) and CrS is
the cryoconcentration cycle.

3.2. Color parameter measurements

Comparisons of the color parameters between cryoconcentrated
milk samples and the control are shown in Table 2. The L* value is a
measure of milk whiteness, a* is the measure of milk color indicating
a tendency to the greenish (if negative) or redness if (if positive),
while the b* value is the measure of milk blueness (if negative) or
yellowness (if positive). As it can be observed on the obtained data,
the milk whiteness increased signicantly between the control (initial

Fig. 1. Casein micelles of the initial skim (9.25% DM) (a) and cryoconcentrated skim milk (25.12% DM) (b) observed in transmission electron microscopy (TEM). The undiluted
cryoconcentrated skim milk was xed in glutaraldehyde at 20 C.

72

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

unconcentrated skim milk to 139 nm in the concentrated up to

25.12% (w/w) total dry matter. In the literature, the study of Karlsson
et al. (2005) reported different results, but the authors concentrated
skim milk by ultraltration. This difference suggests that the action
of cryoconcentration on skim milk casein micelles may be quite different then the effect of pressure driven concentration process such
as ultraltration.
3.4. Cryoconcentration effect on soluble calcium and casein micelle size
distribution

Fig. 3. Particle size distributions of the initial skim milk (C0) and cryoconcentrated
skim milk at different cryoconcentration cycles: C1, C2, and C3: 1st, 2nd, and 3rd
cryoconcentration cycle, respectively.

skim milk) and the cryoconcentrated milk after the 1st cycle 1
(P b 0.001), but no signicant difference was detected between the
whiteness at the three cryoconcentrated cycles. Negative values of the
a* parameter indicate that milks are greener rather than red. All samples
are greener, and no signicant difference was detected between the
cryoconcentrated milks and the control. Positive b* values indicate
yellowness and negative values indicate blueness. The obtained results
in the present study showed that milk yellowness increased as the
cryoconcentration cycle was increased, and all cryoconcentrated
samples were signicantly different from the control (P b 0.001).
These observations are in agreement with the study of Quinones
et al. (1997), who reported that an increase of protein content increases whiteness and yellowness but decreases the greenness of
milk. According to Philips et al. (1995), the L* value of milk has
been demonstrated to have the most positive impact on consumer
appeal. Previous reports have shown that consumers have the
highest appeal for milks with visual properties close to the whole
milk (Owens, Brewer, & Rankin, 2001). Accordingly, we suggest that
the process of skim milk cryoconcentration enables making whiteyellowish milk, which can preserve the similar appearance of nonthermal treated fresh whole milk.
3.3. Casein micelle shape
Observation of the casein micelles from the initial skim milk showed
a roughly spherical shape with various sizes (Fig. 1a). The appearance of
casein micelles of the cryoconcentrated skim milk until 25.12% (w/w)
total dry matter also exhibited nearly spherical shapes with a wide
range of sizes (Fig. 1b). This appearance is in agreement with previous
reports of Srilaorkul et al. (1991) on skim milk concentrated 5-fold
by ultraltration. The structures of casein micelles appear as dark circular close-packed. In the case of our study where skim milk was
cryoconcentration until 25.12% dry matter, there were no signs of
pronounced aggregation or of merged casein micelles. Cryoconcentration
did not result in a signicant reduction of the distance between
casein micelles, which varied from approximately 160 nm in the

Cryoconcentration process of skim milk resulted in a slightly lower

pH (6.5) compared to the non-concentrated skim milk pH (6.7). This
change of pH was accompanied by minerals concentration in the
concentrated milk. However, cryoconcentration efciency decreased
depending on the number of cycles used by 2/3 of Mg, K, P, and less
than 2/3 of Ca and Na in the concentrated part at the rst cycle and
less than 50% and 30% at second and third cycle, respectively (Fig. 2).
The particle size distributions in skim milk concentrates as a function
of the cryoconcentration cycle is shown in Fig. 3. Cryoconcentrated
milk casein micelle size distribution was monomodal, with all the
micelle size ranging from 50 to 350 nm. It has been observed that increasing the concentration cycle tends to reduce the micelle size. The
cryoconcentration process caused a signicant decrease (P b 0.001)
of the average casein micelle size, which ranged from 166.5 nm at
the 1st cryoconcentration cycle to 152 nm at the 2nd cycle and
147.9 nm at the 3rd cycle. The reason for this decrease is likely an
overall contraction, which affects the balance of minerals, especially calcium and phosphor, as shown by the linearly increased (R2 = 0.991)
mineral content in the soluble fraction. The size distribution of casein
micelles obtained from the control (9.25% DM), cycle 1 (14.73% DM),
cycle 2 (21.36% DM), and cycle 3 (25.12% DM) are presented in
Table 3. In this study, casein micelles were classied into 7 classes according to 50 nm increments. Micelles larger than 350 nm in diameter
were grouped together into one class. The obtained results showed
that all the casein micelle size distributions exhibited a maximum in
the diameter range of 100 to 150 nm and a minimum in the diameter
larger than 350 nm. Given that about 60% of the total volume occupied
by the casein micelles has a size of 100200 nm, less than 18% of the volume has a size of 50100 nm and only less than 1% was larger than
350 nm. However, from the control to the cryoconcentration cycle 3,
the proportion of micelles with diameter of 100 to 150 nm increased
from 30.67% to 39.23%. The proportion of micelles smaller than 100 nm
increased also from 10.47% to 18.38%. In the class of micelle size of
300350 nm and N350 nm, the proportion of these micelles decreased
as the cryoconcentration cycle was increased. The reason of these changes may be due to the mineral imbalance, especially Ca and Mg ions, which
occurred after the cryoconcentration fraction was separated (thawed).
The storage of the concentrates for up to 2 weeks at 4 C had no
effect on the size distribution of casein micelles. The reason is likely
that the cryoconcentration process has no signicant effect on the distance between the casein micelles. However, after 3 weeks of storage,
the distribution shifted toward larger micelles with average particle
size of 220 nm (Fig. 4). The increase in particle size from the 3rd week
storage can be explained by the presence of a small number of larger
particles formed from micelle aggregation. Previously, large micelle

Table 3
Size distribution (%) of casein micelles in skim milk and cryoconcentrated milk.
Cryoconcentration cycle

0 (initial)
1st
2nd
3rd

Particle size of casein micelles (nm)

50100

100150

150200

200250

250300

300350

10.477
17.834
16.11
18.388

35.67
36.99
38.147
39.23

25.23
22.412
23.306
22.353

10.893
9.057
9.294
8.607

13.957
10.87
10.802
9.614

2.753
2.005
1.794
1.454

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

Fig. 4. Particle size distributions after 3 weeks of storage: Unconcentrated skim milk (C0),
cryoconcentrated skim milk at different cryoconcentration cycles: C1, C2, and C3: 1st, 2nd,
and 3rd cryoconcentration cycle, respectively.

size indicating aggregation of casein micelles after 6 h has been observed by Liu et al. (2012), who studied the evaporative concentration
of skim milk and its effect on casein micelle hydration, composition,
and distribution. In their work, the skim milk samples were concentrated at 50 C in a rotary evaporator at 355 5 Torr (0.47 bar) for different
times to acquire different concentrations of milk. Moreover, large micelle size was also observed after 8 h in the study of Bienvenue et al.
(2003) on the rheological properties of concentrated skim milk as affected by the soluble minerals and the changes in its viscosity during
storage. The increased micelle size observed in a previous study of
Martin, Williams, and Dunstan (2007) on skim milk powder may have
resulted from the duration it was kept under concentrated conditions.
These results would suggest that the size of casein micelles is affected
by the contraction due to water removal and mineral imbalance.
3.5. Flow behavior
Fig. 5a shows the ow curves recorded at 25 C of the control and
skim milk samples cryoconcentrated from 9.25% to 25.12% total dry

73

matter. At this temperature, up to the 3rd cryoconcentration cycle, similarly to the control sample (initial skim milk), all cryoconcentrated milk
samples exhibited a Newtonian behavior. Comparison between the
control (initial) milk and the samples obtained at the 1st and 2nd
cryoconcentration cycles showed that the apparent viscosity decreased
slowly before to increase with increasing the shear rate. These results
are in agreement with those reported by Sauer et al. (2012) on the caseins concentrated from 2.5% up to 12.5%. However, for the milk concentrate obtained at the 3rd cryoconcentration cycle, the apparent viscosity
decreased slowly until about 25 s1 with increasing the shear rate,
which remained stable up to 100 s1 and decreased rapidly 500 s1.
This behavior is called shear thinning and implies that the liquid has
not true but an apparent viscosity. After 3 weeks of storage, the apparent viscosity of cryoconcentrated skim milk sample of 25.12% total dry
matter (fraction of the 3rd cryoconcentration cycle) increased by nearly
50% (Fig. 5b) and decreased rapidly with increasing shear rate, which
may indicate a fragile (weak) inter-micellar links.
To accurately evaluate the most adapted ow behavior, various
models has been used in the literature to describe the dependence of
viscosity from shear rate: the power law model (Solanki & Rizvi,
2001; Vlez-Ruiz and Barbosa-Cnovas, 1998), the Bingham model
(Bienvenue, et al., 2003), and the HerschelBulkley model Vlez-Ruiz
and Barbosa-Cnovas (1998) are used. In our study, as the total dry
matter content of the cryoconcentrated milk was b 30%, the yield stress
was very small, and the viscosity data for all the cryoconcentrated
milk samples was tted to Power law model and the values of and n
were thus determined. The t for all data sets was very good with a
coefcient of determination N 98% in all cases. From time 0 to week 2,
all cryoconcentrated milk samples exhibited a net Newtonian behavior
(n very close to 1) for all the tested temperatures (5, 20, and 25 C)
(Fig. 6a), and from the week 3 to the week 4 of storage, the n index
decreased with increasing concentration and decreasing working
temperature. In the last week of storage (week 5), at all the tested temperatures, the obtained results showed a non-Newtonian behavior of
the skim milk cryoconcentrated up to 25.12% dry matter (n b 1)
(Fig. 6b). Previously, it has been showed that milk could be regarded
as a Newtonian uid when the total solids content of is 25% Chang
and Hartel (1997).
The ow parameters of the cryoconcentrated skim milk at 3 temperatures and 5 weeks of storage are presented in Tables 4 and 5.
Temperature and concentration effects on the consistency coefcient
were well correlated by an exponential equation with a coefcient of

Fig. 5. (a) Apparent viscosity of control (9.25%), cycle 1 (14.73%), cycle 2 (21.36%), and cycle 3 (25.12%). (b) Apparent viscosity of factor 3 X as function of shear rate for week end 0 (S0),
week end 1 (S1), week end 2 (S2), week end 3 (S3), week end 4 (S4), week end 5 (S5).

74

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

Week: 0, 1 and 2

Flow behavior index

Temperature (C)

Skim milk concentration (%)

Skim milk concentration (%)

Temperature (C)

Week: 5

Flow behavior index

Fig. 6. Combined effect of temperature and concentration on the ow behavior index weeks of storage: (a) weeks 02, (b) week 5.

determination of a least 0.988. However, at the 5th week of storage, the combined effects of [concentration temperature] and
[concentration week] on the consistency showed a highly signicant
difference (P b 0.0001), and no signicant difference (P = 0.0379) was
observed for the combined effect of [temperature week] on the consistency. Different results have been reported by Alvarez de Felipe
et al. (1991) and Vlez-Ruiz and Barbosa-Cnovas (1998) Vlez-Ruiz
and Barbosa-Cnovas (1998) for sweetened condensed and evaporated
milk, respectively, which can be explained by the structural build-up or
gel formation developed during storage (Patil & Patel, 1992).
4. Conclusion
Skim milk was successfully cryoconcentrated by applying the cascade
effect up to 25.12% total dry matter. After 3 cryoconcentration cycles, the
process did not show any signicant effect on the casein inter-micellar
distance. Cryoconcentration resulted in a mineral imbalance, especially
Ca and Mg, and tended to reduce the size of the casein micelles as the

cryoconcentration cycle was increased. This reduction of the casein micelle size was reversible only after 3 weeks of refrigerated storage at
4 C. The ow behavior of the cryoconcentrated skim milk showed a
Newtonian nature at all cryoconcentration cycles. The viscosity increased
with the increasing the cryoconcentration cycle, but a low shear rate had
much more effect on the concentrated milk at higher cycles. After
3 weeks of storage, the cryoconcentrated skim milk of 25.12% total dry
matter showed a non-Newtonian ow behavior only at 5 C, while at
the 5th week of storage, the cryoconcentrated milk with 25.12% total
dry matter content showed a non-Newtonian nature at all the tested
temperatures (5 C, 15 C, and 25 C). The temperature had less effect
on the viscosity versus concentration. Moreover, cryoconcentration of
skim milk signicantly improved its color, as shown by its L* value,
and yielded milk with a whiteness index similar to that of whole
milk. The results obtained on the effects of cryoconcentration on
the physico-chemical properties and rheological behavior during
storage suggest that cryoconcentration cycles 1 and 2 would yield
better results for the use in the manufacture of dairy products.

Table 4
Consistency index (K) of the cryoconcentrated skim milk at different temperatures and storage period (W15 = Weeks 15).
W1

W2
n

K (mPa.s )
5 C
2.3 0.01
4.7 0.03
12.3 0.5
30.3 0.3

K (mPas )
15 C
1.5 0.004
2.8 0.11
6.4 0.3
13.8 0.2

25 C
1.1 0.01
1.9 0.005
4 0.04
6.3 0.2

5 C
2.5 0.2
4.7 0.01
12 0.07
30.9 1.2

W4
n

K (mPas )
15 C
1.4 0.11
2.8 001
6.3 0.07
13.7 0.07

25 C
1.1 0.03
1.7 0.01
3.27 0.02
6.7 0.02

5 C
2.3 0.007
5.1 0.001
13.4 0.1
30.5 1.6

W5
n

K (mPasn)

K (mPas )
15 C
2 0.02
3 0.06
6 0.08
13.9 0.3

25 C
1.0 0.08
1.8 0.01
3.6 0.05
6.8 0.03

5 C
2.40 0.02
5.1 0.23
20.1 0.7
41.5 0.9

15 C
1.59 0.02
4.1 0.01
9.8 0.01
18.6 0.1

25 C
1.1 0.02
1.9 0.17
5.7 0.05
10.6 0.006

5 C
2.4 0.03
5.2 0.003
21.9 0.23
41.4 0.4

15 C
1.3 0.3
5.0 0.001
10.3 0.1
21 0.2

25 C
1.1 0.04
2 0.03
5.9 0.05
11.2 0.23

Data are presented as mean standard deviation.

Table 5
Flow behavior index (n) of the cryoconcentrated skim milk at different temperatures and storage period (W15 = Weeks 15).
W1
C, %
n (dimensionless)
(w/w)
5 C
15 C
9.25
14.73
21.36
25.12

1.012 0.019
1.00 0.012
1.002 0.009
0.978 0.003

25 C

W2

W3

W4

n (dimensionless

n (dimensionless

n (dimensionless

5 C

1.015 0.004 1.015 0.002 1.009 0.001

1.015 0.000 1.021 0.001 1.002 0.002
1.025 0.001 1.011 0.000 0.998 0.001
0.987 0.001 1.011 0.001 0.985 0.001

15 C

25 C

5 C

1.01 0.003
1.009 0.003
1.001 0.001
0.989 0.001

1.01 0.011 1.012 0.007

1.01 0.001 1.005 0.102
1.02 0.000 0.999 0.004
0.988 0.000 0.9 0.002

5 C

W5

15 C

n (dimensionless

15 C

25 C

25 C

5 C

1.00 0.001
1.000 0.000
0.998 0.002
0.989 0.031

1.005 0.000 0.985 0.027 1.00 0.001

1.00 0.000
0.909 0.001
1.01 0.011
0.985 0.029 0.992 0.000 0.998 0.011
0.894 0.001
1.005 0.002 0.962 0.031 0.992 0.001 0.9755 0.002 0.89 0.002
0.915 0.001 0.908 0.008 0.965 0.032 0.965 0.000
0.851 0.000

15 C

25 C

0.905 0.01
0.889 0.000
0.89 0.014
0.868 0.000

0.905 0.005
0.898 0.003
0.889 0.001
0.879 0.001

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

C, % (w/w)
9.25
14.73
21.36
25.12

W3
n

75

76

A. Balde, M. Aider / Innovative Food Science and Emerging Technologies 34 (2016) 6876

Acknowledgments
The authors express their entire gratitude to the Natural Sciences
and Engineering Research Council of Canada (NSERC) for the nancial
support. They heartily thanks Mrs. Diane Gagnon for the excellent
technical support she provided during the realization of this work.
References
Aider, M., de Halleux, D., & Melnikova, I. (2009). Skim milk whey cryoconcentration and
impact on the composition of the concentrated and ice fractions. Food and Bioprocess
Technology, 2(1), 8088.
Aider, M., & Ounis, W. B. (2012). Skim milk cryoconcentration as affected by the thawing
mode: Gravitational vs. microwave-assisted. International Journal of Food Science &
Technology, 47(1), 195202.
Alexander, M., Rojas-Ochoa, L., Leser, M., & Schurtenberger, P. (2002). Structure, dynamics, and optical properties of concentrated milk suspensions: An analogy to hardsphere liquids. Journal of Colloid and Interface Science, 253(1), 3546.
Alvarez de Felipe, A., Melcn, B., & Zapico, J. (1991). Structural changes in sweetened condensed milk during storage: An electron microscopy study. Journal of Dairy Research,
58(03), 337344.
Anema, S. (2008). Effect of milk solids concentration on the gels formed by the acidication of heated pH-adjusted skim milk. Food Chemistry, 108(1), 110118.
Bienvenue, A., Jimenez-Flores, R., & Singh, H. (2003). Rheological properties of concentrated skim milk: Importance of soluble minerals in the changes in viscosity during storage. Journal of Dairy Science, 86(12), 38133821.
Carnovale, V., Britten, M., Couillard, C., & Bazinet, L. (2015). Impact of calcium on the interactions between epigallocatechin-3-gallate and -lactoglobulin. Food Research
International, 77(3), 565571.
Chang, Y. -H., & Hartel, R. W. (1997). Flow properties of freeze-concentrated skim milk.
Journal of Food Engineering, 31(3), 375386.
Dahbi, L., Alexander, M., Trappe, V., Dhont, J., & Schurtenberger, P. (2010). Rheology and
structural arrest of casein suspensions. Journal of Colloid and Interface Science,
342(2), 564570.
De Kruif, C. (1999). Casein micelle interactions. International Dairy Journal, 9(3), 183188.
Di Cesare, L., Cortesi, P., & Martini, E. (1993). Studies on the concentration of model solutions and fruit distillates by freezing-out'. Flssiges Obst, 60(12), 442445.
Fellows, P. (2000). Food processing technology principles and practices (2nd (ed. ed.). Boca
Raton, FL: CRC Press.
Flesland, O. (1995). Freeze concentration by layer crystallization. Drying Technology,
13(89), 17131739.
Ghizzoni, C., Del Popolo, F., & Porretta, S. (1995). Quality of cryoconcentrated red fruit
juices as a function of volatile fraction. Rivista Italiana Eppos, 15, 521.
Gunathilake, M., Dozen, M., Shimmura, K., & Miyawaki, O. (2014). An apparatus for partial
ice-melting to improve yield in progressive freeze-concentration. Journal of Food
Engineering..
Hwang, J., Lee, S., Park, H., Min, S., & Kwak, H. (2007). Comparison of physicochemical and
sensory properties of freeze-concentrated milk with evaporated milk during storage.
ASIAN AUSTRALASIAN JOURNAL OF ANIMAL SCIENCES, 20(2), 273.
IDF, I. (2002). 185. Dtermination de la teneur en azote. Mthode pratique par combustion selon le principe de Dumas. International Dairy Federation, Brussels, Belgium.
Iritani, E., Katagiri, N., Okada, K., Cao, D. -Q., & Kawasaki, K. (2013). Improvement of concentration performance in shaking type of freeze concentration. Separation and
Purication Technology, 120, 445451.

Jeurnink, T. J., & De Kruif, K. G. (1993). Changes in milk on heating: Viscosity measurements. Journal of Dairy Research, 60(02), 139150.
Karlsson, A. O., Ipsen, R., Schrader, K., & Ard, Y. (2005). Relationship between physical
properties of casein micelles and rheology of skim milk concentrate. Journal of
Dairy Science, 88(11), 37843797.
Keshani, S., Luqman Chuah, A., Nourouzi, M., Russly, A., & Jamilah, B. (2010). Optimization
of concentration process on pomelo fruit juice using response surface methodology
(RSM). Int Food Res J, 17, 733742.
Liu, D. Z., Dunstan, D. E., & Martin, G. J. (2012). Evaporative concentration of skimmed
milk: Effect on casein micelle hydration, composition, and size. Food Chemistry,
134(3), 14461452.
Martin, G., Williams, R., & Dunstan, D. (2007). Comparison of casein micelles in raw and
reconstituted skim milk. Journal of Dairy Science, 90(10), 45434551.
Miyawaki, O., Liu, L., Shirai, Y., Sakashita, S., & Kagitani, K. (2005). Tubular ice system for
scale-up of progressive freeze-concentration. Journal of Food Engineering, 69(1),
107113.
Mootse, H., Pisponen, A., Pajumgi, S., Polikarpus, A., Tatar, V., Sats, A., & Poikalainen, V.
(2014). Investigation of casein micelle particle size distribution in raw milk of
Estonian Holstein dairy cows. Agronomy Research, 12(3), 153158.
Morison, K., & Hartel, R. (2006). Evaporation and freeze concentration. Handbook of Food
Engineering (pp. 496550).
Morris, G., Foster, T., & Harding, S. (2000). The effect of the degree of esterication on the
hydrodynamic properties of citrus pectin. Food Hydrocolloids, 14(3), 227235.
Owens, S., Brewer, J., & Rankin, S. (2001). Inuence of bacterial cell population and pH on
the color of nonfat milk. LWT-Food Science and Technology, 34(5), 329333.
Patil, G., & Patel, A. (1992). Viscoelastic properties of sweetened condensed milk as inuenced by storage. Milchwissenschaft (Germany, FR).
Phillips, L. G., Mcgiff, M. L., Barbano, D. M., & Lawless, H. T. (1995). The inuence of fat on
the sensory properties, viscosity, and color of lowfat milk. Journal of Dairy Science,
78(6), 12581266.
Quiones, H. J., Barbano, D. M., & Phillips, L. G. (1997). Inuence of protein standardization by ultraltration on the viscosity, color, and sensory properties of skim and 1%
milk. Journal of Dairy Science, 80(12), 31423151.
Sauer, A., Doehner, I., & Moraru, C. I. (2012). Steady shear rheological properties of micellar casein concentrates obtained by membrane ltration as a function of shear rate,
concentration and temperature. Journal of Dairy Science, 95, 55695579.
Solanki, G., & Rizvi, S. (2001). Physico-chemical properties of skim milk retentates from
microltration. Journal of Dairy Science, 84(11), 23812391.
Srilaorkul, S., Ozimek, L., Ooraikul, B., Hadziyev, D., & Wolfe, F. (1991). Effect of ultraltration of skim milk on casein micelle size distribution in retentate. Journal of Dairy
Science, 74(1), 5057.
Thijssen, H. (1970). Concentration processes for liquid foods containing volatile avours
and aromas. International Journal of Food Science & Technology, 5(3), 211229.
Thijssen, H., & Van Der Malen, B. (1981). Implications on quality of energy savings in the
concentration of foods. Resources and Conservation, 7, 287299.
Vlez-Ruiz, J. F., & Barbosa-Cnovas, G. V. (1998). Rheological properties of concentrated
milk as a function of concentration, temperature and storage time. Journal of Food
Engineering, 35(2), 177190.
Walstra, P. (1979). The voluminosity of bovine casein micelles and some of its implications. Journal of Dairy Research, 46(02), 317323.
Walstra, P., & Jenness, R. (1984). Dairy chemistry & physics. John Wiley & Sons.