JFS C: Food Chemistry

Effect of Solid Fat Content on Structure

in Ice Creams Containing Palm Kernel
C: Food Chemistry

Oil and High-Oleic Sunflower Oil

KRISTINE K. SUNG AND H. DOUGLAS GOFF

ABSTRACT: The development of a structural fat network in ice cream as influenced by the solid:liquid fat ratio at
the time of freezing/whipping was investigated. The solid fat content was varied with blends of a hard fraction of
palm kernel oil (PKO) and high-oleic sunflower oil ranging from 40% to 100% PKO. Fat globule size and adsorbed
protein levels in mix and overrun, fat destabilization, meltdown resistance, and air bubble size in ice cream were
measured. It was found that blends comprising 60% to 80% solid fat produced the highest rates of fat destabilization
that could be described as partial coalescence (as opposed to coalescence), lowest rates of meltdown, and smallest
air bubble sizes. Lower levels of solid fat produced fat destabilization that was better characterized as coalescence,
leading to loss of structural integrity, whereas higher levels of solid fat led to lower levels of fat network formation
and thus also to reduced structural integrity.
Practical Application: Blends of highly saturated palm kernel oil and monounsaturated high-oleic sunflower oil
were used to modify the solid:liquid ratio of fat blends used for ice cream manufacture. Blends that contained 60%
to 80% solid fat at freezing/whipping temperatures produced optimal structures leading to low rates of meltdown.
This provides a useful reference for manufacturers to help in the selection of appropriate fat blends for nondairy-fat
ice cream.
Keywords: ice cream, partial coalescence, vegetable fats

Introduction The objectives of this study were to further understand the influ-

I ce cream is a multiphased system composed of an unfrozen

serum phase, fat globules, air bubbles, and ice crystals (Marshall
and others 2003). The fat structure that exists in ice cream results
ence of variable solid-to-liquid fat compositions on the fat network
structure of ice cream and to further characterize the fat structure
of ice creams made with vegetable fat. The use of vegetable fats
from partial coalescence of the fat during the whipping/freezing in ice creams is very common around the world; however, chang-
step of manufacture. Partial coalescence is an irreversible desta- ing the solid-to-liquid fat ratio at freezing/whipping temperatures
bilization or agglomeration of a dispersed fat phase to form a by selecting fats with differing saturated:unsaturated triglyceride
network of conjoined fat globules (Goff 1997a). The network is held ratios can affect the amount and strength of fat destabilization. A
together by the sintering of the solid crystalline components and hard fraction of palm kernel oil (PKO) and high-oleic sunflower oil
flow of the liquid components of the fat globule with each other, (HOSO) were used to vary the ratio of solid:liquid fat at the temper-
which forms a 3-dimensional network because, unlike coalescence, atures of aging and scraped-surface freezing. Physical parameters
it also maintains the individual identity of the conjoined droplets were used to assess the optimal PKO:HOSO blend, including mix
(Goff 1997b). Partial coalescence as a colloidal phenomenon is dis- and ice cream fat particle size measurements, the quantity of pro-
tinct from coalescence, the complete loss of identity of individ- tein adsorbed to the fat globule interface, the percentage of fat that
ual discrete droplets as they collide and flow into larger droplets, has been destabilized by freezing, the overrun percentage, air bub-
and from flocculation, the weak association of individual discrete ble size distributions, and the rate of meltdown of the frozen ice
droplets in which they do not lose their individual identity and cream.
can be redispersed with minimum input of energy (Dalgleish 2006).
Partial coalescence is a unique phenomenon associated with par- Materials and Methods
tially crystalline emulsion droplets (Vanapalli and Coupland 2001).
Many desirable quality properties of ice cream are governed by the Ice cream preparation
nature of the partially coalesced fat network, including dryness of Ice cream mixes were prepared in 6-L batches with a base formu-
the ice cream at the time of extrusion from the scraped-surface lation of 10% fat, 10% milk solids-not-fat (from skim milk powder;
freezer, slow meltdown, good shape retention, resistance to shrink- Parmalat Canada, Toronto, Ontario, Canada), 12% sucrose (Lantic
ing, and good foam stability. The optimum fraction of solid fat in Sugar, Montreal, Quebec, Canada), 4% 42DE corn syrup solids
the fat phase of ice cream for optimal structure has been suggested (Casco, Toronto, ON, Canada), 0.15% glycerol monostearate (GMS),
to be between 1/3 and 2/3 at 4 to 5 ◦ C (Marshall and others 2003). 0.15% guar gum, 0.015% κ-carrageenan (all from Danisco Inc.,
Toronto, Ontario, Canada), and 63.69% water. The 10% fat com-
MS 20091122 Submitted 11/10/2009, Accepted 1/7/2010. Authors are with ponent for each mix contained a blend of fractionated PKO (ACH
Dept. of Food Science, Univ. of Guelph, Ontario, Canada. Direct inquiries Food Companies Inc., Memphis, Tenn., U.S.A.) and HOSO (Nealan-
to author Goff (E-mail: dgoff@uoguelph.ca).
ders Intl., Toronto, Ontario, Canada), representing the solid and

C274 JOURNAL OF FOOD SCIENCE—Vol. 75, Nr. 3, 2010 

C 2010 Institute of Food Technologists
R

doi: 10.1111/j.1750-3841.2010.01539.x
Further reproduction without permission is prohibited
Structure in ice creams . . .

liquid fat sources, respectively. These fats are of present interest percent of the distribution larger than the 90th percentile (dv ,(0.9) ,
to ice cream manufacturers since PKO provides a natural source μm) of the original mix distribution.
of saturated fat (avoiding hydrogenation) and HOSO provides a
source of monounsaturated oil that is less prone to autoxidation Adsorbed protein analysis
than polyunsaturates. Seven fat blends, ranging from 40% to 100 % Fat droplets were separated from mix emulsions by centrifuga-
(wt/wt) PKO diluted with HOSO, were melted at 40 ◦ C before the tion (Optima LE-80K Ultracentrifuge; Beckman Coulter, Inc., Brea,
addition of the other ingredients to ensure complete mixing of the Calif., U.S.A.) at 40 ◦ C, 12000 rpm for 40 min (Segall and Goff 1999;

C: Food Chemistry
fat blend. Aleong and others 2008). The supernatant layer containing the
Each ice cream mix was initially pasteurized at 75 ◦ C with a hold- protein-coated fat droplets was removed from the centrifuge tube
ing time of 15 min. After pasteurization, a prehomogenization step with a spatula and spread on 70-mm Whatman filter paper (Grade
was carried out with an immersion blender (Silverson L4RT, Sil- nr 2 Qualitative circles) to remove additional serum. The Dumas
verson Machines, Inc., Chesham, Bucks., U.K.) to ensure complete method was used to quantify the total protein/nitrogen content
mixing of the ingredients. The mixes were then homogenized in a in the fat phase, using a FP-528 Protein/Nitrogen Analyzer (LECO
2-stage homogenizer (31MR Laboratory Homogenizer; APV Gaulin Corp., St. Joseph, Mich., U.S.A.). Samples of 0.1 to 0.5 g were placed
Inc., Everett, Mass., U.S.A.), immediately cooled to 4 ◦ C and aged into tin capsules (6.0 mm/15 mm; LECO Corp.) and tightly sealed
for 24 h at 4 ◦ C. Freezing of all the ice cream mixes was performed for analysis. EDTA (nitrogen 9.57 ± 0.026%) was used as the cali-
in a batch freezer (Model 104; Taylor Co., Rockton, Ill., U.S.A.). Ice bration standard. Adsorbed protein was calculated as protein con-
cream mix (1.5-L) was fed into the barrel freezer and whipped and tent of the supernatant layer (mg protein/g fat) divided by specific
frozen until the ice cream reached a target temperature of −5 ◦ C. surface area (m2 /g fat), where the specific surface area was deter-
Whipping continued at this constant temperature until the total mined by integrated light scattering. All treatments were analyzed
time in the freezer reached 10 min. Ice cream was then drawn at in triplicate.
−5 ◦ C and immediately placed into a hardening room at −35 ◦ C for
storage. Mix and ice cream preparation and freezing were carried Overrun
out in triplicate. Subsequent analyses used analysis of variance and Overrun was measured by weighing the frozen ice cream in a
least significant difference testing to show significance among the fixed volume container (227 mL). The overrun percentage was cal-
treatments. culated as (vol. of ice cream − vol. of mix used)/vol. of mix used ×
100%.
Analysis of solid fat content
Measurement of solid fat content in nonemulsified bulk fats Meltdown resistance
blended at the ratios above was carried out using a differential The initial weight of the samples (−20 ◦ C) was measured by re-
scanning calorimeter (DSC Q1000; TA Instruments, New Castle, moving them from the packaging container (120 mL) and then
Del., U.S.A.). Prior to analysis, 3 different calibrations (baseline, cell placing on a mesh grid (2.5 × 2.5 mm) and allowing to stand at am-
constant, and temperature) were carried out at a heating rate of bient temperature (20 ◦ C). The serum passing through the screen
5 ◦ C/min, using indium (m.p. 156.6 ◦ C, Hf = 28.45 J/g) as the stan- was collected and weighed every 10 min for a total time of 90 min.
dard metal. Purified nitrogen (20 psi) was the purge gas for the dry
box. An empty, hermetically sealed aluminum-alodined pan was Air cell distribution
used as the reference sample. Samples of the fat blends (approxi- Using the optical microscopy method developed by Chang and
mately 10 ± 0.1 mg) were weighed into the same aluminum pans Hartel (2002), samples (0.2 to 0.5 mL) were thinly cut with a ra-
and hermetically sealed with an aluminum-alodined cover. Sam- zor blade and placed into a specially made slide with a depression
ples were held isothermally at 80 ◦ C, cooled to −70 ◦ C at a rate of 40 mm deep. The samples were observed on a Linkam cold stage
5 ◦ C/min, held isothermally for 2 min, and then heated to 80 ◦ C at using an optical light microscope (Olympus BX60; Carson Group
a rate of 5 ◦ C/min during which time data were collected. TA Q Se- Inc., Precision Instrument Div., Markham, Ontario, Canada) at ×40
ries Advantage software was used to analyze and plot the thermal and ×100 objective under bright-field illumination. A microscope
data. The onset temperature (T o ), end point temperature (T f ) (the stage micrometer (Fisher Scientific; Zeiss Nr: 474026; Calibration:
point where the extrapolated edge of the exotherm intersects with 5 × 100/100 mm positive; Catalog nr 12-070-63) was viewed with
the baseline), and peak temperatures (temperatures of maximum the same magnifications for image acquisition. Temperature was
heat flow) between T o and T f were determined. The solid fat index slowly raised from −17 to −5 ◦ C (2 ◦ C/min) to allow visualization
(SFI) was calculated as a function of temperature from −20 to 60 ◦ C of air bubbles at the surface of the cover slip while ice crystals re-
by the fraction of the total area under the curve that is passed at any mained beneath.
given temperature. Images were acquired by a video camera (SenSys A99J7015; RS
Images, Photometrics Ltd., Tucson, Ariz., U.S.A.) with a minimum
Fat particle size analysis of 7 images taken, for a total of 250 to 400 air cells to calculate the air
Particle size distributions of fat globules and agglomerated fat in cell diameter, ensuring an accurate representation of the air bub-
the samples were determined by light scattering using a Mastersizer ble size distribution. The images were further processed by an im-
2000 (Malvern Instruments, Malvern, U.K.). Ice cream mix samples age editing computer program (Adobe Photoshop and ImageJ) to
were analyzed immediately after cooling to 4 ◦ C, while the hard- resize the images and to provide the calibration magnification bar
ened ice cream samples (−30 ◦ C) were tempered at 4 ◦ C for 24 h be- onto each image. Sizing of the bubbles was performed using the
fore analysis. Samples were directly diluted (1:1000) in the sample analysis software Object Image 2.10. Measurements of air cell size
chamber with MilliQ water at 16 to 18 ◦ C. Measurements were per- were performed by manually tracing the perimeter of each bubble
formed at ambient temperatures and repeated in triplicate. Specific using a pen tablet. The software automatically calculated the area
surface area (m2 /g fat), volume weighted mean (d4.3 , μm), and sur- of the bubble and the cumulative distribution of the equivalent cir-
face weighted mean (d3.2 , μm) were recorded. The volume fraction cular diameters, which was then characterized by the Logistic Dose
of agglomerated fat in the ice cream samples was calculated as the Response Model, as described by Flores and Goff (1999).

Vol. 75, Nr. 3, 2010—JOURNAL OF FOOD SCIENCE C275

 Structure in ice creams . . .

Results and Discussion Protein adsorption to fat globules

Bolliger and others (2000) presented a direct relationship be-
Solid fat content tween the adsorbed protein content, resulting from its displace-
The melting thermographs of each fat blend utilized in the mix ment by emulsifiers, and the extent of partial coalescence. This
formulations and the pure PKO and HOSO fats were measured. implies that measurement of adsorbed protein may be a good indi-
Peak melting temperature for the HOSO was −7 ◦ C, extending to cator of the potential for partial coalescence to occur during whip-
0 ◦ C, and for PKO was 31 ◦ C. Blends showed 2 peaks at the same ping/freezing. A smaller protein load represents more desorption
C: Food Chemistry

locations with the size of the peak a function of the fat ratio in by the emulsifier, and thus a lower steric stabilization of the droplet,
the blend. Using the melting curves, the SFI over the tempera- which should render it more prone to partial coalescence. The
ture range of −20 to 60 ◦ C was determined. At −5 ◦ C, the lowest amount of protein adsorbed onto the fat globules (mg of protein per
freezing and whipping temperature in the scraped-surface freezer, m2 of surface area) after aging decreased as the fat ratio increased
the SFI ranged from 55 for the 50% PKO sample to 100 for the in PKO from 50% to 100% but was lower at 40% than it was at 50%
100% PKO sample (Figure 1). Although there may be slight differ- (Figure 2). The decreasing adsorbed protein level is due perhaps to
ences, the results from the bulk fat can be applied to the emul- the crystallization of fat at the interface interfering with protein ad-
sified fats in the ice cream mix to give a very good indication of sorption or displacing protein during aging. Typically, the surface
solid fat contents in the mix. The mix was aged for 24 h at 4 ◦ C, protein coverage from an emulsion formed with skim milk pow-
which is adequate time for the fats to attain equilibrium crystal- der is between 6 and 10 mg protein/m2 , with a strong dependence
lization (Berger 1990). Therefore, most of the solid fat at the time on the emulsification conditions (Pelan and others 1997). The GMS
of freezing/whipping was PKO with most of the HOSO remaining emulsifier used here might be expected to displace as much as 30%
liquid. to 50% of the protein (Bolliger and others 2000). De Feijter and oth-
ers (1987) claim that the influence of oil type on the desorption
of protein is negligible and less obvious in comparison to the ef-
Fat globule size distributions in ice cream mix fect of emulsifier type and concentration. However, the effect of
All ice cream mixes displayed a monomodal distribution with a fat crystallization on surface protein adsorption was not discussed.
peak between 0.5 and 0.7 μm ending at 2 to 3 μm, suggesting that Another possibility is that GMS is making a bilayer around the fat
each mix was well homogenized and the fat phase had no influence globule, since distilled saturated monoglycerides like the GMS used
on emulsion droplet formation. in this study have been known to make bilayers in water when

Figure 1 --- The solid fat index at draw

temperature (−5 ◦ C) as a function of
palm kernel oil (PKO) concentration
in a fat blend of PKO and high-oleic
sunflower oil, measured by
differential scanning calorimetry.

Figure 2 --- Protein adsorbed on the fat

globule surface in ice cream mix as a
function of palm kernel oil (PKO)
concentration in a fat blend of PKO
and high-oleic sunflower oil measured
after 24 h of aging at 4 ◦ C.

C276 JOURNAL OF FOOD SCIENCE—Vol. 75, Nr. 3, 2010

Structure in ice creams . . .

present in high concentrations (Krog 1997). Bilayers might be ex- the issue associated with nonspherical aggregates of fat being char-
pected to result in different protein desorption behaviors com- acterized by a mean size. Similarly, in this study, frozen ice creams
pared to monolayers. Regardless of the mechanism, the decreasing had larger fat globule distributions than their mixes and showed
amount of adsorbed protein with increasing solid fat content needs typical bimodal peaks with the smaller-size peak being normally
to be considered when looking at the effect of solid fat content on distributed but the larger-size peak being irregular. As the PKO
partial coalescence, since adsorbed protein also affects partial co- concentration, and hence solid fat, increased the amount of fat
alescence directly making it difficult to sort out which parameter destabilization decreased (Table 1); and 70% to 80% PKO showed

C: Food Chemistry
might have more influence on partial coalescence. a significantly larger percent of particles >dv ,(0.9) than 90% to 100%
PKO, despite the reduced adsorbed protein loads in these high PKO
Overrun samples discussed above (Figure 2). This agrees with our under-
standing of the importance of an optimum ratio of solid fat, which
The measurement of overrun from the batch freezer provides
includes some small portion of liquid fat, on the development of
a measurement of the foamability of the mix. The overrun of the
partial coalescence and suggests this solid fat ratio is more impor-
ice cream at −5 ◦ C when extruded from the batch freezer after
tant, in this instance, than adsorbed protein content. Ice creams
10 min of whipping ranged from 60% to 80% and was significantly
with 40% to 60% PKO had greater percent of particles >dv ,(0.9) than
influenced by the fat fraction composition. Higher overrun percent-
70% to 80% PKO. As the solid fat content decreased, it would be ex-
ages were achieved as the PKO concentration increased (Figure 3),
pected that coalescence of the oil-rich droplets would be enhanced
suggesting higher foamability and foam stabilization of the mixes
and structural partial coalescence levels would be reduced. Fur-
within the batch freezer. The 100% PKO ice cream had the largest
ther interpretation of the light scattering data supports this con-
increase in volume due to incorporation of air. Air bubbles are sta-
cept. The sample with 40% PKO showed a significantly lower d3,2
bilized by adsorption of the proteins present in the concentrated
value than the samples with 50% to 60% PKO (Table 1), which
serum phase onto the serum–air bubble interface and by the ad-
reflects the differing nature of the distribution (Figure 4). Once
sorption of the fat phase onto the air interface (Goff and others
the process of coalescence begins to dominate over partial coa-
1999). However, overrun itself is not necessarily related to fat struc-
lescence, the nature of the particles measured by light scattering
ture since a higher level of fat adsorption at air interfaces may de-
changes. Partial coalescence is characterized as being distinctly bi-
crease overrun but result in a more stable foam to resist meltdown
modal, representing both individual droplets and fat aggregates (as
and structural collapse (Koxholt and others 2001; Muse and Hartel
shown by the sample with 90% PKO in Figure 4). However, as co-
2004). The lower overrun values in the 40% PKO ice cream could
alescence dominates (as shown by the sample with 40% PKO in
be due to oil wetting or spreading. Fat globules can spread at the
Figure 4) the peak of original mix droplets progressively disappears
air interface (Goff and Jordan 1989; Goff and others 1999), which
and the coalesced droplets appear at progressively larger sizes.
may have been facilitated by the lack of solid fat (high percentage of
HOSO) and agitation due to the shear forces imparted by the dash-
ers during freezing. The spreading of oil can lead to the thinning
of the lamella between bubbles (Hotrum and others 2004), causing Table 1 --- Effect of increasing palm kernel oil (PKO) con-
centration on fat particle size parameters in frozen ice
destabilization and collapse of the air phase. In turn, this may cause creams measured by integrated light scattering.
an effect similar to churning with a weaker structure as the end
% PKO (wt/wt)A % Particles > d v,(0.9) (μm)∗, ∗∗ d 3,2 (μm)∗
result.
40 85.7 ± 0.7a 2.6 ± 0.07b
50 92.1 ± 1.3a 4.4 ± 0.3a
Fat globule size distributions in ice cream 60 88.7 ± 1.3a 5.5 ± 1.8a
Previous studies have described the 2nd peak in fat globule size 70 82.5 ± 9.1b 3.2 ± 1.5b
distributions of ice cream to result from fat destabilization during 80 81.2 ± 1.8b 2.4 ± 0.2b
90 66.0 ± 6.3c 1.2 ± 0.1c
freezing in which larger particles have developed from the initially
100 55.0 ± 1.9c 0.9 ± 0.1c
small fat globules of the mix. This 2nd peak is typically character- A
PKO diluted with high-oleic sunflower oil to 100% (wt/wt).
ized by a measure of the movement from the size of the discrete ∗ Each listed value is the average of triplicate sets of measurements ± standard
droplets in the mix to larger agglomerates of fat, for example, as ∗∗ deviation (n = 3).
Percentage of particles with mean diameters greater than the cumulative vol-
percent of particles in the ice cream >dv ,(0.9) in the mix (Gelin and ume distribution at the 90th percentile of the mix (dv ,(0.9) , μm).
others 1994; Bolliger and others 2000). This value helps to alleviate a,b,c
values with the same letter in the same column do not differ (α = 0.05).

Figure 3 --- Overrun (%) in ice cream as

a function of increasing palm kernel
oil (PKO) concentration in a fat blend
of PKO and high-oleic sunflower oil at
constant freezing time (10 min) and
draw temperature (−5 ◦ C).

Vol. 75, Nr. 3, 2010—JOURNAL OF FOOD SCIENCE C277

 Structure in ice creams . . .

Thus, light scattering may be able to distinguish between the var- Air bubble sizes
ious forms of fat destabilization if more than one parameter is used The air cells of the ice creams were imaged to obtain their size
in interpretation. distributions and their role in stabilizing or destabilizing the ice
creams (Table 2). The mean air cell diameters for all of the ice
Meltdown resistance creams ranged between 11.3 and 23.3 μm. There was a significant
The melting and shape-retention properties of ice cream are per- difference in bubble size amongst the vegetable fat ratios, with
haps the most significant and observable ramification of structure 60% to 80% PKO producing smaller bubbles than either 40% to
C: Food Chemistry

development (Koxholt and others 2001; Muse and Hartel 2004). 50% PKO or 90% PKO. It has been generally accepted that the ad-
Shape retention results from the interactions between fat glob- dition of air enhances rapid flocculation and fat globule coales-
ules with each other and between fat globules and air, as the ice cence (Walstra and Oortwijn 1982; Goff and Jordan 1989), which,
crystals melt away. Lowest melting rates were seen at 50% to 80% in turn, adsorbs back to air interfaces and helps stabilize them.
PKO (Figure 5). At higher levels of solid fat, the ice creams melted Thus the optimal level of solid fat content to produce the high-
at a faster rate indicating lack of network structure development est levels of meltdown resistance were also seen to produce the
(Vanapalli and Coupland 2001). This correlates well with the fat smallest air bubbles which, in turn, would also aid in structure de-
structure measurements by light scattering. Likewise, at the low- velopment. Bubble size at 100% PKO did not fit that pattern, per-
est level of solid fat, 40% PKO, melting rate was again higher. As haps due to a significant lack of fat structure formation around the
discussed above, the high levels of fat destabilization at low lev- bubble as discussed above. At the lower levels of solid fat, spread-
els of solid fat likely resulted from enhanced coalescence and lim- ing and oil migration of the fat agglomerates around the air bub-
ited structural partial coalescence and the resulting structure was bles during freezing could once again be the cause of air bubble
thus less able to hold its shape as the ice crystals melted. Another instability leading to larger sizes. Hotrum and others (2005) found
factor to consider is that the ice creams with higher PKO contents that shear-induced and surface-mediated partial coalescence fur-
had higher overrun, which may also have increased the rate of melt ther increased fat globule aggregation in emulsions with the incor-
(Muse and Hartel 2004). Further collapse of the ice cream foam oc- poration of an air phase. Fat globules adhere to the interface of the
curs during melt as the serum phase begins to be diluted by the bubbles while shear forces simultaneously deform the bubbles. If a
melting ice crystals. Nevertheless, it is likely that the solid fat con- high level of oil spreading causes the bubbles to rupture, the par-
tent was more dominant in affecting meltdown rate since the over- tially coalesced fat surrounding the bubble can collide with other
run from the lowest to the highest level of PKO only varied from 60% agglomerates to form a larger coalesced fat network (Hotrum and
to 80%. others 2005). Eventually the coalesced fat network becomes so large
that a maximum volume of air that can be stabilized is reached. Fur-
ther shear will lead to a collapse of the stabilized air phase, which,
in turn, will lead to poor shape retention and poor overall quality of
ice cream.

Table 2 --- Mean equivalent diameter (μm) of air cell dis-

tributions of ice creams with increasing palm kernel oil
(PKO) concentration.
% PKO Mean∗ Median Max Min Std. dev.
(wt/wt)A (μm) (μm) (μm) (μm) (μm) Count
b
40 23.3 18.4 94.0 5.6 15.4 239
50 19.2b 16.7 53.1 3.2 10.3 239
60 11.3a 9.6 46.1 1.4 6.0 249
70 13.2b 10.8 60.8 3.8 8.6 246
80 13.9a,d 11.2 60.3 3.2 8.8 239
90 22.5b,c 17.3 73.1 5.4 14.4 242
100 11.9a 10.0 61.6 3.5 7.6 257
A
PKO diluted with high-oleic sunflower oil to 100% (wt/wt).
Figure 4 --- Fat particle size distributions in ice creams at ∗
Each listed value is the use of triplicate images (n = 3) with a total cumulating
40% and 90% palm kernel oil (PKO) in a fat blend of PKO count >220 air cells.
and high-oleic sunflower oil.
a,b,c
Values with the same letter do not differ (α = 0.05).

Figure 5 --- Melt rate of ice cream as a

function of increasing palm kernel oil
(PKO) concentration in a fat blend of
PKO and high-oleic sunflower oil.

C278 JOURNAL OF FOOD SCIENCE—Vol. 75, Nr. 3, 2010

Structure in ice creams . . .

Conclusions Acknowledgments
P KO and HOSO were used in this study to systematically vary the Author Sung was supported by a scholarship from the Canadian
solid:liquid fat ratio in ice creams. It was seen that blends com- Dairy Commission to complete this research.
prising 60% to 80% solid fat produced highest rates of fat destabi-
References
lization that could be described as partial coalescence (as opposed Aleong JM, Frochot S, Goff HD. 2008. Ice recrystallization inhibition in ice cream by
to coalescence), lowest levels of meltdown, and smallest air bub- propylene glycol monostearate. J Food Sci 73(9):463–8.
ble sizes. Although protein adsorption and overrun were also in- Berger KG. 1990. Ice cream. In: Larsson K, Friberg SE, editors. Food emulsions. 2nd
ed. New York: Marcel Dekker, Inc. p 367–444.

C: Food Chemistry
fluenced, with intermediate levels of each seen at these solid fat Bolliger S, Kornburst B, Goff HD, Tharp BW, Windhab EJ. 2000. Influence of emulsi-
contents, the greatest impact on fat network structuring resulted fiers on ice cream produced by conventional freezing and low-temperature extru-
sion processing. Int Dairy J 10:497–504.
from optimization of solid fat content. During the process of par- Chang Y, Hartel RW. 2002. Measurement of air cell distributions in dairy foams. Int
tial coalescence of fat during whipping/freezing of ice cream mix, Dairy J 12:463–72.
Dalgleish DG. 2006. Food emulsions—their structures and structure-forming proper-
it is believed that the fat crystals within the fat globules distort ties. Food Hydrocolloids 20:415–22.
the surface of the globule thus promoting collisions in the high De Feijter JA, Benjamins J, Tamboer M. 1987. Adsorption displacement of proteins by
surfactants in oil-in-water emulsions. Colloids Surf 27:243–66.
shear conditions in the freezer. As globules collide, the oil phase Flores AA, Goff HD. 1999. Ice crystal size distributions in dynamically frozen model
increases adhesion between the globules, resulting in the build- solutions and ice cream as affected by stabilizers. J Dairy Sci 82:1399–407.
Gelin JL, Poyen L, Courthaudon JL, Le Meste M, Lorient D. 1994. Structural changes
up of a 3-dimensional network of fat. Although the process of in oil-in-water emulsions during the manufacture of ice cream. Food Hydrocolloids
partial coalescence is well known, the research reported here en- 8:299–308.
hances our understanding of this process by correlating structure Goff HD. 1997a. Colloidal aspects of ice cream—a review. Int Dairy J 7:363–73.
Goff HD. 1997b. Instability and partial coalescence in whippable dairy emulsions.
formation to a specific content of crystalline fat within fat glob- J Dairy Sci 80:2620–30.
ules. In addition to solid fat content, the nature of the fat crys- Goff HD, Jordan WK. 1989. Action of emulsifiers in promoting fat destabilization dur-
ing the manufacture of ice cream. J Dairy Sci 72:18–29.
tals within the droplet may influence partial coalescence formation Goff HD, Verespej E, Smith AK. 1999. A study of fat and air structures in ice cream. Int
and could vary amongst different natural sources of fat. Thus, the Dairy J 9:817–29.
Hotrum NE, Cohen Stuart MA, van Vliet T, van Aken GA. 2004. Spreading of partially
specific results for PKO seen here may not extrapolate directly to crystallized oil droplets on an air/water interface. Colloids Surf A 240:83–92.
other sources of natural saturated fat, such as coconut oil; never- Hotrum NE, Cohen Stuart MA, van Vliet T, Avino SF, van Aken GA. 2005. Elucidating
the relationship between the spreading coefficient, surface-mediated partial coa-
theless, the colloidal principles would be expected to remain the lescence and the whipping time of artificial cream. Colloids Surf A 260:71–8.
same. Koxholt MMR, Eisenmann B, Hinrichs J. 2001. Effect of the fat globule sizes on the
meltdown of ice cream. J Dairy Sci 84:31–7.
This study should be of great commercial interest to the ice Krog N. 1997. Food emulsifiers. In: Gunstone FD, editor. Lipid technologies and ap-
cream industry. In North America, there is a growing trend toward plications. New York: Marcel Dekker, Inc. p 521–6.
Marshall RT, Goff HD, Hartel RW. 2003. Ice cream. 6th ed. New York: Springer.
the manufacture of frozen desserts in which nondairy fats are re- Muse MR, Hartel RW. 2004. Ice cream structural elements that affect melting rate and
placing milk fat, similar to the trends that have occurred in Europe hardness. J Dairy Sci 87:1–10.
Pelan BMC, Watts KM, Campbell IJ, Lips A. 1997. The stability of aerated milk protein
in the last decades. These results will help manufacturers to opti- emulsions in the presence of small molecule surfactants. J Dairy Sci 80:2631–8.
mize the blend of fats for both functionality and cost. There is also Segall KI, Goff HD. 1999. Determination of protein surface concentration for emul-
sions containing a partially crystalline dispersed phase. Food Hydrocolloids
a desire to reduce saturated fats in fat blends used for ice cream and 13:291–7.
the results presented here will help in developing strategies for re- Vanapalli SA, Coupland JN. 2001. Emulsions under shear—the formation and proper-
ties of partially coalesced lipid structures. Food Hydrocolloids 15:507–12.
placement of the structural role of saturated fat while maintaining Walstra P, Oortwijn H. 1982. The membranes of recombined fat globules. 3. Mode of
optimal quality. This might allow for the addition of highly unsatu- formation. Neth Milk Dairy J 36:103–13.
rated oils, like omega-3 fish oils, in ice cream formulations.

Vol. 75, Nr. 3, 2010—JOURNAL OF FOOD SCIENCE C279