Accepted Manuscript

Effect of solubilised carbon dioxide at low partial pressure oncrystallisation behaviour, microstructure and texture of anhydrousmilk fat

Tuyen Truong, Martin Palmer, Nidhi Bansal, Bhesh Bhandari

PII: S0963-9969(17)30087-XDOI: doi: 10.1016/j.foodres.2017.02.022Reference: FRIN 6605

To appear in: Food Research International

Received date: 29 November 2016Revised date: 22 February 2017Accepted date: 26 February 2017

Please cite this article as: Tuyen Truong, Martin Palmer, Nidhi Bansal, Bhesh Bhandari, Effect of solubilised carbon dioxide at low partial pressure on crystallisation behaviour,microstructure and texture of anhydrous milk fat. The address for the corresponding authorwas captured as affiliation for all authors. Please check if appropriate. Frin(2017), doi:10.1016/j.foodres.2017.02.022

This is a PDF file of an unedited manuscript that has been accepted for publication. Asa service to our customers we are providing this early version of the manuscript. Themanuscript will undergo copyediting, typesetting, and review of the resulting proof beforeit is published in its final form. Please note that during the production process errors maybe discovered which could affect the content, and all legal disclaimers that apply to thejournal pertain.

ACC

EPTE

D M

ANU

SCR

IPT

1

Effect of solubilised carbon dioxide at low partial pressure on

crystallisation behaviour, microstructure and texture of anhydrous milk fat

Tuyen Truong, Martin Palmer, Nidhi Bansal, Bhesh Bhandari*

ARC Dairy Innovation Hub

School of Agriculture and Food Sciences, the University of Queensland, QLD 4072,

Australia

Abstract

The crystallisation and melting behaviour, fat polymorphs, microstructure and texture of

anhydrous milk fat (AMF) was investigated in the presence of dissolved CO2 (0 – 2000 ppm)

under two crystallising conditions (non-isothermal versus isothermal). CO2 was found to

induce higher onset crystallisation temperature during cooling from 35 to 5oC at 0.5

oC min

-1.

X-ray scattering analysis showed that, in the presence of dissolved CO2, this rapid

crystallisation caused the formation of unstable, polymorph fat crystals. For milk fat

crystallised under isothermal condition at 25oC for 48h, dissolved CO2 improved solid fat

content, slightly depressed melting temperature and exhibited a sharper melting peak.

Microstructure of AMF visualised by Polarised light microscopy of crystallised AMF showed

that increasing dissolved CO2 concentration was associated with smaller crystal size and

greater crystal number. The bulk properties of the fat appeared to mirror the microstructural

differences, in that the texture of CO2-treated AMF was harder under isothermal condition

but became softer than untreated AMF under cooling condition. The results of this study are

of significance in understanding how CO2 treatment might be used to modulate the

crystallisation behaviour of milkfat and thereby the structural development and physical

functionality of fat-containing dairy products.

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

2

Keywords: AMF; milkfat; crystallisation; CO2; microstructure; texture

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

3

1. Introduction

Milk fat is an important determinant of quality and functionality in milk and dairy products.

Milk lipids comprise 98% triacylglycerols (TAG) and have a widely varied fatty acid

composition with more than 400 species (Christie, 1995). The melting point range of milk

lipids is correspondingly wide, varying in temperature from -40 to 40oC and commonly

exhibiting three overlapping endotherms. These correspond to three incompletely miscible

fractions of milk fat (low-, middle-, and high-melting point fractions). In its solid state, milk

fat consists of a mixture of crystals with three main polymorphic forms (, ’, and ), of

which ’ polymorph is dominating (Sato, 2001). The type of fat crystal network, which is

dependent on the particular combination of fat polymorphs and the crystal habit (size and

shape), determines the microstructure of milk fat, which in turn, helps to determine the

physical functionality of dairy products such as butter, cheese and whipped cream (Danmark

& Bagger, 1989; Goff, 2002; Lopez, Camier, & Gassi, 2007; Moens, Masum, & Dewettinck,

2016). Therefore, being able to understand and control the crystallisation behaviour of milk

fat is essential in manufacturing fat-structured dairy products of consistent quality.

For bulk fat, crystallisation behaviour is governed by fat composition, cooling rate and

processing conditions. Fat crystallisation can also be influenced by dissolved or suspended

entities present in the fat. Most crystallisation occurs in complex food systems with more

than one ingredient. Ingredients may be present as natural admixtures or impurities or they

may be deliberately incorporated as additives in food formulations. Such ingredients may

have a considerable effect on nucleation and crystal growth mechanisms and kinetics, thus

impacting crystallization structure and properties of fat or other components (Hartel, 2001).

In fact, impurities are commonly found to inhibit or facilitate nucleation and crystallisation,

leading to the use of compounds like glycerol monooleate or sucrose esters, which are added

to control crystallization in some fat systems. The presence of these additives acts to

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

4

influence the inter-solubility of low and high melting point TAGs, thereby influencing the

crystallisation behaviour of the high melting point TAGs.

Carbon dioxide (CO2) is a hydrophobic gas that is soluble in milkfat and other lipids. This

gas is used in many modified atmospheric packaging systems for fruits, vegetables, drinks,

meat and cheese. This gas also has the highest solubility in both muscle and fat tissues among

other gases (Gill, 1988; Mitz, 1979) and higher solubility in water than nitrogen (N2) (50–70

times less) or oxygen (O2) (25–35 times less soluble) (Jakobsen & Bertelsen, 2006; Jakobsen,

Jensen, & Risbo, 2009; Mitz, 1979). The gas solubility coefficients of N2, O2 and CO2 in

hydrogenated lard at 65oC are reported as 0.0844, 0.145 and 0.920 vol /vol fat, respectively

(Formo, 1979). Our previous study (Truong, Palmer, Bansal, & Bhandari, 2017) showed that

CO2 is highly soluble in anhydrous milk fat (AMF) with a solubility coefficient of 1648 ppm

kg fat-1

atm-1

. The dissolution behaviour appeared to be dependent on both intrinsic (partial

pressure and temperature) and extrinsic variables (fatty acid composition and physical state

of the AMF).

When CO2 is in its supercritical phase (above 31.1oC and 7.38 MPa), it can be employed in

various food applications such as removal of undesirable components in food (i.e. cholesterol

in milk fat, milk powder and alcohol in cider), encapsulation of liquids, extraction of

carotenoids and natural complex (i.e. phenols and tannins) (Bradley, 1989; Brunner, 2005;

Chitra, Deb, & Mishra, 2015; Murga, Ruiz, Beltran, & Cabezas, 2000; Sun & Temelli, 2006).

Being soluble in lipids, CO2 can take part in the crystallisation process, in turn, affecting the

product property. It was reported that supercritical CO2 impacted on the crystallisation and

melting properties of polymers such as polypropylene and polystyrene (Cao et al., 2014). In

food-related system, recently Uniliver has patented a new process known as “Supercritical

Melt Micronisation (ScMM)” which is a spray crystallisation technology in which fat melt is

saturated with CO2 at high pressure and atomised through a nozzle (Bezemer, Huizinga, &

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

5

Malseen, 2012). Due to the rapid pressure drop at the nozzle, the CO2 evaporates causing

rapid super-cooling, crystallisation and expansion of the fat particles. The extremely porous

nature of the fat particles allows blending it with oil to make a stable fat spread. This is due to

the adsorption of oil into the pores of fat. This is a high pressure process and requires

specialised equipment. To the best of our knowledge, there is no published research on the

influence of CO2 at low pressure on the crystallisation behaviour, microstructure and

functionality of an edible fat system.

This study aimed to investigate effect of dissolved CO2 at low partial pressure (0 – 1 atm) on

the crystallisation, microstructural and textural properties of AMF, with a view to exploring

the industrial potential of this technique to manipulate the functional properties of fat-

structured dairy products.

2. Materials and Methods

2.1 Materials

Anhydrous milk fat was purchased from Rogers & Company Foods Pty Ltd (Hampton East,

Victoria, Australia). It was stored in cold room (4oC) until use. Dry ice, the solid form of

CO2, was supplied in pelleted form by University of Queensland (Brisbane, Australia). Dry

ice was stored in foam box at -80oC to prevent sublimation. Weight of individual dry ice

pellets was approximately 1 -3 g. The transformation of solid CO2 (dry ice) to gas phaseCO2

happens at -78.5oC, which is its sublimation point at atmospheric pressure (Perry & Green,

2007).

2.2 Dissolution of CO2 into AMF

CO2 was dissolved into AMF using a home-made apparatus, which is fitted with a digital

manometer (DC-400, GTS Gauges Transmitters Switches, Midvale, Western Australia)

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

6

connected to the screw cap of a Schott bottle (0.627 L) (Truong et al., 2017). When the dry

ice was placed into the container and the screw cap immediately closed, the dry ice

sublimated in the closed system. In presence of the AMF, difference in gaseous

concentrations between the headspace and the sample is a driving force for a gas transfer

until the equilibrium state is obtained. Since this apparatus was operational in a static and

batch modes, the known amount of initial dry ice and the partial pressure reading allowed

calculation of the dissolved CO2 concentration into the milk fat sample (in continuous mode

without opening the system) as follows:

s

headspaceCO

eqCORTW

MwCOVpC

2

.2

2

1000 [Eq. 1]

where CCO2 eq. is the equilibrium amount of dissolved CO2 (ppm by mass) taking into account

any related dissociation and hydration products, pCO2 (pCO2 initial – pCO2 final) is the

difference in partial pressure drop due to CO2 gas adsorption into sample, Vheadspace is the

headspace gas volume (L), MW is the molecular weight of CO2 (44.01 g mol-1

) and Ws is the

weight of the milk fat (kg).

2.3 Experimental procedure

Figure 1 depicts the experimental procedure. The AMF was completely melted at 70oC for 1

h and cooled to 35oC. At 35

oC, dry ice was added into liquid AMF at three concentrations (0,

1000, and 2000 mg kg-1

) and equilibrated for 24 h in an incubator, also at 35oC. Dissolution

of CO2 into AMF was monitored by changes in partial pressure, as described above.

Investigation of the effects of dissolved CO2 on the physical properties and crystallisation

behaviour of the AMF was then undertaken under two crystallising conditions:

(a) The first scheme involved non-isothermal (cooling) crystallisation of the AMF, from

35 to 5oC at a cooling rate of 0.5

oC min

-1. Nucleation and fat crystal growth of the

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

7

control and CO2-treated AMF were examined in the first 60 min of solidification

process by polarised light microscopy. In parallel, 40 mL of each sample was sub-

divided into a plastic container and allowed to cool to 5oC. The sub-samples were

stored at 5oC for 48h and then measured for CO2 retention, hardness and thermal

properties.

(b) In the second scheme, the control and CO2-treated AMF was cooled from 35 to 28oC

at a cooling rate of 0.5oC min

-1 and isothermally crystallised at 28

oC. Examination of

the crystal habit during solidification process for 60 min was undertaken at 28oC,

using polarised light microscopy. Sub-samples (40 mL each) were immediately

transferred into plastic container and allowed to cool to room temperature (25oC).

Similar to the first scheme, after 48 h of storage, measurements of CO2 retention,

hardness, and thermal properties were also carried out.

2.4 Measurements

2.4.1 Absorption and retention of CO2 into AMF

The CO2 concentration dissolving into liquid AMF after 24 h of equilibration at 35oC was

calculated according to Eq. 1. The CO2 retention level in (semi-) solid AMF after storage for

48 h at either 5 or 25oC was measured using the titration method described by Jakobsen and

Bertelsen (2006). In this method, two 100 mL Buchner flasks, one containing 0.5M H2SO4 (5

mL) and the other containing a standardised Ba(OH)2 solution 0.1N, were connected by

neoprene tubing. Approximately 2 g of the sample was placed into the flask containing

H2SO4 solution. Upon equilibration for 18 – 24 h, CO2 was evolved and reacted with the

standardised Ba(OH)2, yielding precipitated BaCO3. Standard HCl solution (0.1 M) was used

to titrate the residual amount of Ba(OH)2 with phenolphthalein as the indicator.

2.4.2 Differential Scanning Calorimetry (DSC)

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

8

Crystallisation and thermal properties of the AMF were measured using a Differential

Scanning Calorimeter (DSC1 STARe System, Mettler-Toledo, Schwerzenbach, Switzerland).

Calibration were undertaken using an indium standard (melting point = 156.66°C, ΔH

melting = 28.41 J g-1

). Approximately 10-15 mg of AMF was placed in a 40 μL aluminium

pan, which was then hermetically sealed (an empty pan was used as a reference).

Crystallisation properties of the sample were investigated in cooling mode (#1 scheme) while

heating profile was monitored for both schemes. The sample in #1 scheme was cooled in the

DSC pan from 35 to 5oC at a cooling rate of 0.5

oC min

-1. Regarding thermal properties,

depending on the investigation scheme, the DSC pan was heated from either 5oC or 25

oC to

60oC at a heating rate of 5

oC min

-1 to measure endpoint of melting (TM-endset) and melting

enthalpy (HM). Solid fat content (SFC) was calculated based on the relative proportion of

total melting enthalpy to that of reference (bulk milk fats solidified at -70oC). STAR

e

Excellence Software (Mettler-Toledo, Schwerzenbach, Switzerland) was used to analyse

thermographs. All measurements were performed at least in duplicate.

2.4.3 Small and wide angle X-ray scattering (SAXS/WAXS)

Structure and polymorphs of fat crystals in solid state at 5oC (#1 scheme) were examined

using a SAXSess mc2 combined SAXS/WAXS spectrometer (Anton Paar GmbH, Graz,

Austria) as described elsewhere (Truong, Morgan, Bansal, Palmer, & Bhandari, 2015).

Briefly, the sample was placed in a paste cell (CAT. 17233, Anton Paar GmbH, Graz,

Austria), which was previously cooled at 5oC. The paste cell was then inserted to a pre-

chilled sample holder at 5oC. Equilibration was undertaken in the sample holder at 5

oC for 30

min before measurement at the same temperature for 20 min. Peakfit software (SeaSolve

Software Inc., San Jose, CA, US) was used to analyse the SAXS/WAXS peak position and

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

9

peak area (integrated intensity) as detailed by Lopez, Lavigne, Lesieur, Bourgaux, and

Ollivon (2001a).

2.4.4 Polarised light microscopy (PLM)

Nucleation and growth of fat crystals were observed using PLM. An aliquot of AMF (10 L)

was pipetted onto a microscope slide, which was then placed onto a LTS120 sample chamber

connected to a PE-95 Linkpad temperature controller (Linkam Scientific Instruments, UK).

This Peltier controlled system covers working temperature from -40oC to 120

oC with 0.1

oC

temperature control and stability. The pipette tip, microscopic slide and cover slip were all

pre-warmed at the investigated temperature (i.e. 35 or 28oC) before sampling. Observation of

microstructure of AMF was conducted using an Olympus CX41 microscope (Tokyo, Japan)

at 4 magnification. Image analysis was undertaken by ImageJ software (Schneider,

Rasband, & Eliceiri, 2012).

2.4.5 Texture analysis

Hardness of the AMF, expressed as the maximum force (N), was measured using a texture

analyser (TA.TX2 Stable Micro Systems, UK) (Wright, Scanlon, Hartel, & Marangoni,

2001). In a compression test mode, the 2 mm cylindrical probe was set to travel a distance of

10 mm with test speed of 1 mm s-1

. Each measurement was triplicated. The longer-term effect

of CO2 on hardness of the AMF was also examined after 9 days of storage.

2.4.6 Statistical analysis

All measurements were done at least in duplicate. Statistical analyses of the data were

conducted using the statistical package MINITAB ® Released 16 (Minitab Co., US). One-

way ANOVA and Tukey’s LSD was employed to determine significant differences of

treatment means at P<0.05.

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

10

3. Results and Discussion

3.1 CO2 absorption and retention

Table 1 summarises the dependence of dissolved CO2 concentrations on initial CO2

concentrations, showing that CO2 solubilities increased with increasing initial CO2

concentrations. CO2 was highly dissolved in liquid AMF at 35oC/24 h with 953 and 1887

ppm for 1000 and 2000 ppm of initial concentrations, respectively. This agreed well with our

observation in the previous study on solubility of CO2 in AMF (Truong et al., 2017). When

AMF was cooled either from 35 to 5oC (#1 scheme) or 35

oC to 25

oC (#2 scheme) and then

stored at the final temperatures for 48 h, retention of CO2 was still more than 60% at all

conditions (Table 1). Accordingly, any observed modifications in physical and textural

properties can then most likely be ascribed to the presence of dissolved CO2.

3.2 Crystallisation and thermal properties

Crystallisation and thermal properties of AMF was measured using DSC. For the #1 scheme,

all AMF samples showed similar DSC cooling patterns with two main crystallisation peaks

(Figure 2A), indicating the crystallisation of high- and low-melting point TAGs. It is clear

that CO2 promoted the nucleation process in the first 60 min of solidification during cooling

from 35 to 5oC (Figure 2A). The onset crystallisation temperature of AMF under CO2

dissolution (23.8 and 25.3oC) was significantly higher than that of original AMF (19.2

oC)

(Table 2). Crystallisation enthalpies also slightly increased with higher amounts of dissolved

CO2. This increasing trend in crystallisation parameters could be explained by CO2 acting as

a template for crystallisation of AMF. In other studies, both hydrophobic and non-

hydrophobic additives have been reported to promote lipid crystallisation (Smith, Bhaggan,

Talbot, & van Malssen, 2011; Yoshikawa, Kida, & Sato, 2014). Hydrophobic additives

having similar configuration chemically or structurally (monoacylglycerols, monopalmitin) to

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

11

lipids can act as seeding nuclei for lipid crystallisation (Fredrick et al., 2013; Verstringe,

Danthine, Blecker, Depypere, & Dewettinck, 2013) while non-hydrophobic additives

(inorganic particles: carbon nanotube, graphite, and talc; organic molecules: theobromine,

terephthalic acid) can interact with lipid molecules via hydrophilic interactions (Yoshikawa et

al., 2014). Being a soluble gas, it is possible that CO2 could promote crystal nucleation during

cooling. However, it seems that the effect of supercooling was dominating in the #1 scheme

as the differences in melting properties of the CO2 dissolved AMF after cooling and storing at

5oC for 48 h were not statistically significant (Table 2). DSC thermal profiles of both non-

treated and CO2-treated AMF in the #1 scheme were similar with three major melting peaks

at 5-10oC, 10-23

oC, and 23-39

oC. In contrast, influence of absorbed CO2 on melting

properties of AMF was more pronounced in #2 scheme, which had no supercooling. While

the low melting peaks (range of 25-30oC) of the three samples remained unchanged, the

endset melting point was slightly depressed (Table 2). The main melting peak (range of 30 –

43oC) at 2000 ppm addition was shifted to a lower temperature (34

oC) and sharper than those

of the lower CO2 concentration (37oC) (Figure 2C), indicating melting of more uniform fat

crystals. Furthermore, melting enthalpy (indicating for relative crystallinity or solid fat

content) was enhanced with increasing dissolved CO2 concentration (Table 2). This result

agrees with observations on crystallisation of polymers that dissolved CO2 facilitates mobility

of polymer chains and enhances crystal growth, resulting in increased crystallinity of

polymers (Cao et al., 2014; Lei, Ohyabu, Sato, Inomata, & Smith, 2007; Nofar, Zhu, & Park,

2012).

3.3 Fat polymorphs

As dairy products are generally stored under refrigeration, further characterisation of fat

crystals formed in the #1 scheme was performed by X-ray scattering at 5oC. As shown in

Figure 3, X-ray scattering patterns of TAG in all samples showed a typical profile of milk fat

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

12

(Truong et al., 2015). In the SAXS region, double- (2L) and triple-chain length (3L)

structures were noted for TAG longitudinal packings. In the WAXS region, ’ and fat

polymorphs were detected, representing TAG lateral packings. Noticeably, when AMF was

crystallised in the presence of CO2, AMF exhibited a peak at 4.15 Å, representing fat

crystals. Previous study reported that formation of less close-packed fat crystals indicates

occurrence history of rapid crystallisation and/or a higher number of fat crystals (Lopez,

Lavigne, Lesieur, Bourgaux, & Ollivon, 2001b). In this study, this is consistent with the

earlier crystallisation temperatures measured by DSC (Table 2) and higher number of

crystals seen by microscopy as discussed later (section 3.4). In addition, as the proportion of

phase represented about one-third (34 = 36%) of the total crystalline phases (Table 3), it

would be expected that melting temperatures and melting enthalpies of CO2 –treated AMF

would be lower compared to the untreated sample. However, no association between crystal

polymorph type and melting properties was observed. This might be ascribed to the

dominating effects of the ’ and crystal forms in melting characteristics. In other words, the

increase in crystal proportion was insufficient to induce a noticeable decrease in melting

properties. Previous study on crystallisation of other lipid materials also reported insignificant

correlation between crystal structures and melting behaviour under some specific conditions

(Frydenberg, Hammershoj, Andersen, & Wiking, 2013).

3.3.1 Microstructure

Figure 4 shows the evolution of AMF microstructure during cooling from 35 to 5oC at 0.5

oC

min-1 (

#1 scheme). The first visible fat crystals in CO2-treated AMF were detected at 22.5oC,

which was earlier than that of the untreated AMF (20oC). This observation was very close to

values obtained from DSC at the same cooling rate (Table 2). Quantitative analysis of PLM

images on solid fraction area and number of crystals (Figure 5 A&B), demonstrated that

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

13

crystal nucleation happened faster and a larger number of fat crystals formed in the CO2 -

treated AMF. This was consistent with X-ray scattering, which showed crystal formation in

the presence of CO2. Figure 5A shows that, solidification was faster in the CO2 -treated

AMF, with the maximum solid fraction at 15oC while the untreated AMF had the highest

solid fraction at the end of crystallisation process (5oC). The number of fat crystals reached a

peak at 15oC and decreased, presumably due to crystal-crystal interaction to form bigger

crystals at lower temperatures (Figure 5B). At 5oC, the solid fraction and number of crystals

of CO2 –treated AMF were only slightly higher than the values for the control sample (Figure

5 A&B). This indicates that induction of crystallisation was promoted by dissolved CO2

during cooling. However, in this case the effect of dissolved CO2 was not carried over to the

end of the cooling regime as the microstructure of all AMF samples was quite comparable at

5oC. This means the effect of supercooling overrules the effect of dissolved CO2 obtained

during nucleation. When AMF was crystallised isothermally at 28oC (#2 scheme), crystal

morphology was similar for all treatments, with well-formed spherulites present in all

samples. However, there were marked differences in fat crystal size and number of fat

crystals. For each time point, smaller crystals and more crystals were observed with

increasing dissolved CO2 concentration (Figure 6). Quantification of solid fraction and

particle counts based on PLM images showed that crystal nucleation was induced much faster

on 2000 ppm CO2 addition (Figure 5 C&D). For example, after 20 min of crystallisation at

28oC, the solid fraction in 2000 ppm CO2-treated AMF (4%) was 2 to 3-fold greater than that

at the lower CO2 concentration or control (Figure 5C). Correspondingly, AMF treated with

2000 ppm CO2 yielded about 2 to 3 times more crystals than that of the 1000 ppm treatment

and control sample at 20 min of crystallisation (Figure 5D). The increasing trend of both

solid fraction and crystal counts in 2000 ppm added AMF continued until the end of

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

14

experiment duration (90 min). This implies that CO2 also promotes crystal growth, even over

longer crystallisation times.

3.3.2 Hardness

The physical and functional properties of lipids are greatly governed by their

crystallisation/melting behaviour. For both schemes investigated , i.e. cooling and isothermal

crystallisation, it is obvious that presence of dissolved CO2 impacted on the

crystallisation/melting profiles, the fat polymorphs and the microstructure of fat crystal

network formed. These differences resulted in changes in hardness of AMF at specific

conditions as illustrated in Figure 7. Hardness of AMF stored at 25oC significantly increased

with increasing CO2 concentration (P < 0.05). This result is in accordance with the DSC and

PLM results, showing that AMF exhibited a sharper melting profile, greater degree of

crystallisation, smaller crystal sizes and larger number of crystals in the presence of dissolved

CO2. Interestingly, under refrigeration conditions, softer texture was observed in the CO2 -

treated AMF (P < 0.05) even though its melting properties were very similar to those of the

untreated AMF. In fact, the larger number of small fat crystals in the CO2-treated AMF

provided more contact points in the fat phase, the rigidity of the fat crystal network should be

enhanced. Thus the texture of CO2-treated AMF should be harder in this case (as

demonstrated at 25oC). The opposing result suggests that the hardness of AMF depends on

many interrelated factors. Possible explanations for softer texture of CO2-treated AMF fat

crystal network at 5oC may have been due to (i) the presence of the larger proportion of

crystals induced by CO2 addition (Figure 3); (ii) an important role of interaction strength

between structural elements of fat crystal network (i.e. fat crystal/fat crystal interactions) and

(iii) the interference of embedded gas bubbles in the rigidity of the fat crystal network.

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

15

Figure 7 also shows that there was a slight increase in hardness of all AMF samples when

measuring texture again after 9 days of storage. However, these changes are not statistically

significant different (P > 0.05) between day 2 and day 9 of storage duration (Figure 7). The

changes in hardness of AMF in longer term is related to post-crystallisation phenomenon,

which might be attributable to polymorphic transformations, evolution of new crystals and

presence of solid bridges between structural components (Johansson & Bergenstahl, 1995).

4. Conclusions

This study demonstrates the impact of dissolved CO2 at low partial pressure on

crystallisation, melting behaviour, microstructural and textural properties of AMF under two

contrasting (isothermal vs. non-isothermal) crystallisation regimes. CO2 promoted crystal

nucleation and growth under both crystallisation conditions, as well as an accumulation of the

metastable fat crystal polymorph. CO2 treatment also resulted in a reduction in crystal size

and increase in total number of fat crystals, which in turn, affecting the hardness of AMF at

both 25 and 5oC. These initial results suggest that there may be some potential for industrial

use of CO2 to modulate the physical functionality of milkfat. Further studies are now needed

to investigate whether these effects are also apparent in more complex systems, such as butter

and dairy blends. It will also be critical to understand the optimum amounts of CO2 required

to generate the desired effects as well as the possibility of any concomitant changes to

sensory properties and chemical stability.

Conflict of interest

There are no conflicts of interest regarding this paper.

Acknowledgements

This research was supported under Australian Research Council's Industrial Transformation

Research Program (ITRP) funding scheme (project number IH120100005). The ARC Dairy

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

16

Innovation Hub is a collaboration between The University of Melbourne, The University of

Queensland and Dairy Innovation Australia Ltd.

References

Bezemer, A. J., Huizinga, H., & Malseen, K. F. V. (2012). E. P. Office.

Bradley, R. L. (1989). Removal of Cholesterol from Milk-Fat Using Supercritical Carbon-Dioxide. Journal of

Dairy Science, 72(10), 2834-2840. doi: DOI 10.3168/jds.S0022-0302(89)79429-7

Brunner, G. (2005). Supercritical fluids: technology and application to food processing. Journal of Food

Engineering, 67(1-2), 21-33. doi: 10.1016/j.jfoodeng.2004.05.060

Cao, X. W., Lin, H., Luo, J., He, G. J., Zhang, Y. J., & Xu, B. P. (2014). Comparison Study on CO2-Promoted

Crystallization and Melting Behavior of Polypropylene: Homopolymer and Copolymers. Journal of

Macromolecular Science Part B-Physics, 53(11), 1726-1738.

Chitra, J., Deb, S., & Mishra, H. N. (2015). Selective fractionation of cholesterol from whole milk powder:

optimisation of supercritical process conditions. International Journal of Food Science and

Technology, 50(11), 2467-2474. doi: 10.1111/ijfs.12914

Christie, W. W. (1995). Composition and Structure of Milk Lipids. In P. F. Fox (Ed.), Advanced Dairy

Chemistry Volum 2: Lipids (2nd ed., Vol. 2, pp. 1-36). London: Chapman & Hall.

Danmark, H., & Bagger, L. H. (1989). Effects of Temperature Treatment of Sweet Cream on Physical-

Properties of Butter .1. Factors Affecting Hardness and Consistency. Milchwissenschaft-Milk Science

International, 44(3), 156-160.

Formo, M. W. (1979). Chapter 3: Physical properties of fats and fatty acids. In D. Swern (Ed.), Bailey’s

Industrial Oil and Fat Products (4th ed., Vol. 1, pp. 194-195, 216). New York: John Wiley & Sons.

Fredrick, E., Moens, K., Heyman, B., Fischer, S., Van der Meeren, P., & Dewettinck, K. (2013).

Monoacylglycerols in dairy recombined cream: I. The effect on milk fat crystallization. Food Research

International, 51(2), 892-898.

Frydenberg, R. P., Hammershoj, M., Andersen, U., & Wiking, L. (2013). Ultrasonication Affects Crystallization

Mechanisms and Kinetics of Anhydrous Milk Fat. Crystal Growth & Design, 13(12), 5375-5382.

Gill, C. O. (1988). The Solubility of Carbon-Dioxide in Meat. Meat Science, 22(1), 65-71. doi: Doi

10.1016/0309-1740(88)90027-7

Goff, H. D. (2002). Formation and stabilisation of structure in ice-cream and related products. Current Opinion

in Colloid & Interface Science, 7(5-6), 432-437. doi: Pii S1359-0294(02)00076-6

Hartel, R. W. (2001). Crystallization in foods. Gaithersburg, Md.: Aspen Publishers.

Jakobsen, M., & Bertelsen, G. (2006). Solubility of carbon dioxide in fat and muscle tissue. Journal of Muscle

Foods, 17(1), 9-19. doi: DOI 10.1111/j.1745-4573.2006.00029.x

Jakobsen, M., Jensen, P. N., & Risbo, J. (2009). Assessment of carbon dioxide solubility coefficients for

semihard cheeses: the effect of temperature and fat content. European Food Research and Technology,

229(2), 287-294. doi: DOI 10.1007/s00217-009-1059-3

Johansson, D., & Bergenstahl, B. (1995). Sintering of Fat Crystal Networks in Oil during Post-Crystallization

Processes. Journal of the American Oil Chemists Society, 72(8), 911-920. doi: Doi

10.1007/Bf02542069

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

17

Lei, Z. G., Ohyabu, H., Sato, Y., Inomata, H., & Smith, R. L. (2007). Solubility, swelling degree and

crystallinity of carbon dioxide-polypropylene system. Journal of Supercritical Fluids, 40(3), 452-461.

doi: DOI 10.1016/j.supflu.2006.07.016

Lopez, C., Camier, B., & Gassi, J. Y. (2007). Development of the milk fat microstructure during the

manufacture and ripening of Emmental cheese observed by confocal laser scanning microscopy.

International Dairy Journal, 17(3), 235-247. doi: 10.1016/j.idairyj.2005.12.015

Lopez, C., Lavigne, F., Lesieur, P., Bourgaux, C., & Ollivon, M. (2001a). Thermal and Structural Behavior of

Milk Fat. 1. Unstable Species of Anhydrous Milk Fat. Journal of Dairy Science, 84, 756-766.

Lopez, C., Lavigne, F., Lesieur, P., Bourgaux, C., & Ollivon, M. (2001b). Thermal and structural behavior of

milk fat. 1. Unstable species of anhydrous milk fat. J Dairy Sci, 84(4), 756-766.

Mitz, M. A. (1979). CO2 biodynamics: a new concept of cellular control. Journal of theoretical biology,. 80,

537-551.

Moens, K., Masum, A. K. M., & Dewettinck, K. (2016). Tempering of dairy emulsions: Partial coalescence and

whipping properties. International Dairy Journal, 56, 92-100. doi: 10.1016/j.idairyj.2016.01.007

Murga, R., Ruiz, R., Beltran, S., & Cabezas, J. L. (2000). Extraction of natural complex phenols and tannins

from grape seeds by using supercritical mixtures of carbon dioxide and alcohol. Journal of Agricultural

and Food Chemistry, 48(8), 3408-3412. doi: 10.1021/jf9912506

Nofar, M., Zhu, W., & Park, C. B. (2012). Effect of dissolved CO2 on the crystallization behavior of linear and

branched PLA. Polymer, 53(15), 3341-3353. doi: DOI 10.1016/j.polymer.2012.04.054

Perry, R. H., & Green, D. W. (2007). Perry's Chemical Engineers' Handbook: McGraw-Hill.

Sato, K. (2001). Crystallization behavior of fats and lipids - a review. Chemical Engineering Science, 56, 2255-

2265.

Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH Image to ImageJ: 25 years of image analysis.

Nature Methods, 9(7), 671-675. doi: 10.1038/nmeth.2089

Smith, K. W., Bhaggan, K., Talbot, G., & van Malssen, K. F. (2011). Crystallization of Fats: Influence of Minor

Components and Additives. Journal of the American Oil Chemists Society, 88(8), 1085-1101.

Sun, M., & Temelli, F. (2006). Supercritical carbon dioxide extraction of carotenoids from carrot using canola

oil as a continuous co-solvent. Journal of Supercritical Fluids, 37(3), 397-408. doi:

10.1016/j.supflu.2006.01.008

Truong, T., Morgan, G. P., Bansal, N., Palmer, M., & Bhandari, B. (2015). Crystal structures and morphologies

of fractionated milk fat in nanoemulsions. Food Chemistry, 171, 157-167. doi: DOI

10.1016/j.foodchem.2014.08.113

Truong, T., Palmer, M., Bansal, N., & Bhandari, B. (2017). Investigation of solubility of carbon dioxide in milk

fat by lab-scale manometric method. Food Chemistry, (submitted).

Verstringe, S., Danthine, S., Blecker, C., Depypere, F., & Dewettinck, K. (2013). Influence of monopalmitin on

the isothermal crystallization mechanism of palm oil. Food Research International, 51(1), 344-353.

Wright, A. J., Scanlon, M. G., Hartel, R. W., & Marangoni, A. G. (2001). Rheological properties of milkfat and

butter. Journal of Food Science, 66(8), 1056-1071.

Yoshikawa, S., Kida, H., & Sato, K. (2014). Promotional effects of new types of additives on lipid

crystallization. Journal of Oleo Science, 63, 333- 345.

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

18

Fig. 1

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

19

Fig. 2

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

20

Fig. 3

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

21

Fig. 4

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

22

Fig. 5

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

23

Fig. 6

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

24

Fig. 7

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

25

Figure 1. Flow chart of experimental plan

Figure 2. Representative cooling and heating profiles of AMF subjected two crystallisation

conditions: cooling from 35 to 5oC at 0.5

oC min

-1 (A), stored at 5

oC for 48h and heated from 5 to 60

oC

at 5oC min

-1 (B - #1 scheme); heating only from 25 to 60

oC at 5

oC min

-1 upon solidification of AMF at

25oC/48 h (C- #2 scheme).

Figure 3. Longitudinal and lateral packings of the AMF at various dissolved CO2 concentrations at

5oC upon 48 h of storage (#1 scheme).

Figure 4. Polarised light micrographs of AMF during cooling from 35 to 5oC at cooling rate of 0.5

oC

min-1

.

Figure 5. Percentage of area fraction (degree of crystallisation) and number of particle counts (crystal

number) in AMF as a function of dissolved CO2 levels during cooling (A & B) and isothermally

crystallised at 28oC (C & D) conditions.

Figure 6. Effect of dissolved CO2 concentration on milk fat crystal habit upon isothermally

crystallisation at 28oC for first 90 min.

Figure 7. Impact of dissolved CO2 concentration on AMF hardness upon storage for 2 and 9 days at

the two crystallising conditions; for each storage temperature regime (i.e. 25 and 5oC), same letter

denotes insignificant difference (P <0.05) among three levels of carbonation.

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

26

Table 1. Absorption and retention of CO2 (mg kg-1

) in the AMF

Initial CO2

conc.

Dissolved CO2 at

35oC/ 24 h

CO2 retention at

5oC/48 h

CO2 retention at

25oC/48 h

0 2228 2228 2228

1000 95312 82320 63757

2000 188727 137930 120035

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

27

Table 2. Effect of dissolved CO2 on crystallisation and thermal properties of AMF under the two crystallisation

conditions

Initial CO2

level

TC-onset Crystallization

enthalpy

TM-endset Melting

enthalpy

mg kg-1 o

C J g-1 o

C J g-1

#1 scheme (supercooling)

0 19.2±0.4 b 34.6±5.6

a 38.5±0.1

a 104.1±3.6

a

1000 23.8±1.3 a 37.7±1.3

a 39.8±1.0

a 110.3±2.6

a

2000 25.3±1.6 a 40.0±0.1

a 40.4±1.5

a 111.3±6.6

a

#2 scheme (isothermal)

0 n/a n/a 43.7±0.2 a 33.0±1.1

b

1000 n/a n/a 43.7±0.0 a 31.8±0.9

b

2000 n/a n/a 43.4±0.1 b 43.0±1.0

a

For each scheme, the superscript shared the same letter at the same column had no significant difference (P <

0.05)

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

28

Table 3. Effect of dissolved CO2 concentrations on relative proportion of TAG longitudinal and lateral

structures of the AMF

CO2 level 3L 2L Proportion of polymorph forms

under total WAXS scattered peaks,

% SAXS

peak

% area SAXS

peak

% area ’

0 ppm 64.18 Å 17.97 41.63 Å 82.03 0.0 80.2 19.9

1000 ppm 65.59 Å 33.91 41.54 Å 66.09 36.2 50.9 12.9

2000 ppm 63.84 Å 28.36 41.93 Å 71.64 34.1 53.7 12.2

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

29

Graphical abstract

ACCEPTED MANUSCRIPT

ACC

EPTE

D M

ANU

SCR

IPT

30

HIGHLIGHTS

Dissolved CO2 induced higher onset crystallisation upon cooling of AMF

Dissolved CO2 caused formation of unstable, polymorph fat crystals

Carbonation on AMF modified thermal properties and texture of AMF

CO2 treatment resulted in a reduction in crystal size in microstructure of AMF

Dissolved CO2 can be used to modulate the physical functionality of milkfat

ACCEPTED MANUSCRIPT