Journal of Colloid and Interface Science 460 (2015) 247–257

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

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

Long-term stability of crystal-stabilized water-in-oil emulsions

Supratim Ghosh a,⇑, Mamata Pradhan b, Tejas Patel b, Samira Haj-shafiei b, Dérick Rousseau b,⇑
a
Department of Food and Bioproduct Sciences, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
b
Department of Chemistry and Biology, Ryerson University, Toronto, Ontario, Canada

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o a b s t r a c t

Article history: The impact of cooling rate and mixing on the long-term kinetic stability of wax-stabilized water-in-oil
Received 4 June 2015 emulsions was investigated. Four cooling/mixing protocols were investigated: cooling from 45 °C to
Revised 28 August 2015 either 25 °C or 4 °C with/without stirring and two cooling rates – slow (1 °C/min) and fast (5 °C/min).
Accepted 31 August 2015
The sedimentation behaviour of the emulsions was significantly affected by cooling protocol. Stirring
Available online 1 September 2015
was critical to the stability of all emulsions, with statically-cooled (no stirring) emulsions suffering from
extensive aqueous phase separation. Emulsions stirred while cooling showed sedimentation of a waxy
Keywords:
emulsion layer leaving a clear oil layer at the top, with a smaller separation and droplet size distribution
Emulsion stability
Pickering
at 4 °C compared to 25 °C, indicating the importance of the amount of crystallized wax on emulsion sta-
Network bility. Light microscopy revealed that crystallized wax appeared both on the droplet surface and in the
Wax crystallization continuous phase, suggesting that stirring ensured dispersibility of the water droplets during cooling
Interfacial crystallization as the wax was crystallizing. Wax crystallization on the droplet surface provided stability against droplet
Coalescence coalescence while continuous phase wax crystals minimized inter-droplet collisions. The key novel
Sedimentation aspect of this research is in the simplicity to tailor the spatial distribution of wax crystals, i.e., either at
Microstructure the droplet surface or in the continuous phase via use of a surfactant and judicious stirring and/or cooling.
Contact angle
Knowledge gained from this research can be applied to develop strategies for long-term storage stability
of crystal-stabilized W/O emulsions.
Ó 2015 Elsevier Inc. All rights reserved.

⇑ Corresponding authors at: Department of Food and Bioproduct Sciences, 1. Introduction

University of Saskatchewan, Saskatoon, Saskatchewan S7J 2X3, Canada (S. Ghosh).
Department of Chemistry and Biology, Ryerson University, Toronto, Ontario M5B
2K3, Canada (D. Rousseau).
Emulsions in products such as cosmetics, pharmaceuticals and
E-mail addresses: supratim.ghosh@usask.ca (S. Ghosh), rousseau@ryerson.ca (D. foods rely on the presence of crystallized species such as polymers,
Rousseau). paraffins or triacylglycerols for long-term stability [1–6]. Water-in-

http://dx.doi.org/10.1016/j.jcis.2015.08.074
0021-9797/Ó 2015 Elsevier Inc. All rights reserved.
248 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
oil (W/O) emulsions used in such applications must satisfy differ-
ent criteria including: (i) stability against physical breakdown (par-
ticularly once in the hands of the consumer), (ii) resistance to
changes in environmental conditions (temperature, vibration/
shaking) and (iii) lack of change in quality attributes (texture,
etc.) over time. Key factors include the composition and volume
fraction of the continuous and dispersed phases, the mean droplet
size and distribution of the dispersed phase, the presence and con-
centration of stabilizing agents (surfactants, thickeners or crystals)
and how the emulsion is processed and stored [7]. Finally, the
interactions between the various components in an emulsion will
also affect their stability [8].
Emulsion processing requires that the droplets be well-
dispersed within the continuous phase, which is normally achieved
via valve homogenization or high-shear mixing. In many thickened
or solid-like emulsions, subsequent gelation or crystallization of the
continuous phase helps to keep the dispersed droplets in place. Yet,
little is known regarding how to tailor the spatial distribution of
crystals in oil-continuous emulsions to assist in retarding physical
breakdown. If crystals are adsorbed to the oil–water interface, they
create a solid shell around individual dispersed droplets which pre-
vents droplet–droplet coalescence [9,10]. Such surface-active spe-
cies are usually surfactants that solidify directly at the interface
or adsorb onto the droplet surface [3]. As they are surface-
inactive, many crystalline species such as wax or triacylglycerols
will form a continuous phase network that envelops the dispersed
aqueous phase thereby limiting droplet–droplet collisions [8].
We have previously shown that the surfactant glycerol mono-
oleate can promote the heterogeneous interfacial nucleation of
hydrogenated canola oil (a surface-inactive fat) at the water dro-
plet surface in W/O emulsions [11]. We also showed that the addi-
tion of a dispersed aqueous phase significantly altered the
rheological properties of W/O emulsions [12], as has also been
reported by others [5]. Visintin et al. characterized emulsion rheol-
ogy, and showed that presence of a dispersed aqueous phase was a
key contributor to emulsion elasticity [13]. Finally, Hodge and
Rousseau showed that the addition of crystallized fat prior to, or
after, homogenization (and subsequently quench-crystallized)
enhanced the stability of surfactant-stabilized W/O emulsions,
with the latter conferring greater stability [14,15].
This paper discusses the impact of cooling rate and stirring on
the long-term properties of wax-stabilized W/O emulsions. The
goal was to investigate how the presence of a dispersed aqueous
phase impacted emulsion macroscopic phase separation, wax crys-
tal morphology, solid wax content and droplet size distribution. In
particular, we were interested in showing how it is possible to use
minimal amounts of surfactant to generate W/O emulsions with
long-term stability against physical breakdown. This was achieved
through the use of a low concentration of wax in the continuous oil
phase that was suitably crystallized. Importantly, we showed that
the optimal stabilization of model W/O emulsions depends on the
concerted role of interfacial and network wax crystallization, with
the combination of stirring and rapid cooling from the melt confer-
ring the highest kinetic stability.
The results from this study are relevant to processed products
that exist as oil-continuous emulsions and may serve as a guide
for the judicious formulation of emulsions where long-term kinetic
stability is desired.
2. Materials and methods
2.1. Materials
All emulsions consisted of a 20 wt% aqueous phase and a 80 wt%
oil phase containing 5 wt% paraffin wax, mineral oil and sufficient
surfactant to aid in emulsification, yet not keep the emulsion stable
for any length of time. The light mineral oil was obtained from
Fisher Scientific (Nepean, ON, Canada). The stated maximum kine-
matic viscosity at 40 °C was 33.5 mPa s and its density (q) was
0.84 g/ml at 25 °C. A 3.5 wt% sodium chloride (Fisher Scientific,
Nepean, ON, Canada) in de-ionized water (resistivity
P18.0 MOhm cm) solution was used as the dispersed aqueous
phase (q = 1.02 g/ml). The surfactant used was glycerol monoole-
ate (GMO) (Dimodan MO 90Ò, Danisco, New Century, KS, USA).
Dimodan MO 90 is a food-grade low-HLB (hydrophilic/lipophilic
balance) emulsifier composed of >92% glycerol monooleate, with
the rest consisting of small amounts of glycerol, glycerol dioleate
and glycerol trioleate. A highly-refined paraffin wax (IGI-1242;
m.p. 56.7–58.9 °C) was obtained from The International Group,
Inc. (Toronto, ON, Canada). Its n-paraffin composition provided
by the manufacturer is given in Fig. 1.
2.2. Emulsion preparation
A schematic of the emulsion preparation protocol is shown in
Fig. 2. The oil phase contained 0.05 wt% GMO as emulsifier and
5 wt% paraffin wax. This level of emulsifier allowed for the forma-
tion of dispersed droplets, but did not prevent them from measur-
ably coalescing within a few days in the absence of stabilizing wax.
All emulsions were formulated to initially consist of dispersed
aqueous droplets with a mean diameter of
30 lm with a maxi-
mum size of
300 lm, which permitted direct comparison
between the different W/O emulsions. The aqueous phase con-
tained 3.5 wt% NaCl, which was necessary to prevent undue forma-
tion of oil droplets in the water phase (hence double emulsion
formation), which we ascribed to charge generation by the intense
shearing action of the homogenizer [16]. Pre-mixing was per-
formed by slow addition of the aqueous phase (9 ml/min) to the
oil phase in a beaker at 45 °C via magnetic stirring at 500 rpm for
15 min. The pre-mix was quickly transferred to a valve homoge-
nizer (APV-1000, APV, Albertslund, Denmark) preheated to 45 °C
and homogenized at 500 psi for 4 cycles. Afterwards,
100 ml
samples were collected in 250 ml glass beakers (diameter: 6.5 cm
and height: 9 cm) and cooled via one of the four methods denoted
at ‘quench-static’, ‘quench-stirred’, ‘slow-static’ or ‘slow-stirred’.
For quench cooling, samples were transferred to a 0 °C waterbath
whereas slow-cooled samples were placed in a temperature-
controlled waterbath (Model 1157P, VWR International, Missis-
sauga, ON) at 45 °C and cooled at a rate of 1 °C/min. Stirring during
cooling was performed using a magnetic stirrer at 300 rpm. All
samples were prepared in triplicate. Emulsion temperatures during
the cooling regimes were measured along the centreline of the bea-
Fig. 1. n-Paraffin compositional analysis of the solid wax sample. Analysis is
performed by high temperature gas chromatography yielding a n-paraffin distri-
bution of the wax from C20 to C90.
 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
Fig. 2. Schematic of the emulsion preparation and stabilization protocol.
ker using an auto logger thermocouple (K type with thermometer
HH2002AL, Omega, Montréal, QC, Canada) and downloaded to a
computer using an IR-sensor connection (HHP-2000DL, Omega,
Montréal, QC, Canada). Cooling was stopped when the centre of
the sample in the beaker had reached the desired temperature
(25 °C or 4 °C).
2.3. Emulsion storage
All samples were stored at either room temperature (25 ± 1 °C)
or at refrigeration temperature (4 ± 1 °C) for
1.5 years (78 wks).
At set intervals, samples were analyzed via sedimentation beha-
viour, microscopy, solid wax content (SWC) and water droplet size
determination.
2.4. Sedimentation
Emulsion sedimentation behaviour was recorded visually and
with a digital camera for 78 wks. The height of the sedimented
emulsion layer was calculated from the digital images using Micro-
soft Office Picture Manager (Microsoft Canada, Mississauga, ON,
Canada). Emulsion destabilization was measured based on the
percent-emulsified layer (Femulsion) retained during storage:
hemulsion
F emulsion ¼
htotal
where hemulsion and htotal are the heights of the sedimented emulsion
layer and the initial sample, respectively.
2.5. Microscopy
Samples were taken from the storage beakers with a glass rod,
placed on a viewing slide, covered with a cover slip and observed
with an inverted Zeiss Axiovert 200M light microscope (Zeiss
Inc., Toronto, ON, Canada) equipped with a temperature-
controlled stage (model TSA02i with STC200 controller, Instec,
Boulder, CO, USA) pre-set at the desired temperature. The viewing
slides and the cover slips were cooled to the sample temperature
prior to sampling. A 20 objective lens was used to record sample
microstructure under DIC (Differential Interference Contrast) and
PLM (Polarized Light Microscopy) modes.
249
2.6. Droplet size distribution
The dispersed droplet size distribution of the emulsions was
determined at 25 °C and 4 °C using a Bruker Minispec Mq pulsed
field gradient nuclear magnetic resonance (pfg-NMR) unit (Bruker
Canada, Milton, ON, Canada) that allows unimodal characterization
of emulsion droplet size distributions via restricted diffusion mea-
surement [17]. The temperature inside the NMR sample chamber
was controlled by an external waterbath pre-set at specific tem-
peratures. The Minispec Mq NMR software version was 2.58 revi-
sion 12/NT/XP (Bruker Biospin GmbH, Rheinstetten, Germany)
and the water droplet size application was v5.2 revision 4a. The
pulsed gradient separation and number of pulse widths were
210 ms and 8 ms, respectively. The oil suppression delay was
85 ms and the magnet gradient strength was 2 T/m. The pfg-
NMR field gradient strength was calibrated with CuSO4-doped
water (diffusion coefficient = 2.3  109 m2 s1 at 25 °C). Emulsion
samples (height = 1 cm) in glass tubes (ID = 0.8 cm, L = 20 cm) were
placed in the NMR unit and their droplet size distribution was
determined by d3,3 (volume-weighted geometric mean diameter)
values and the breadth of distribution, r (geometric standard devi-
ation). Destabilized water droplets that phase-separated from the
emulsion were calculated as free water and used as an indicator
of emulsion (in)stability.
2.7. Solid wax content (SWC)
Emulsions were transferred into NMR tubes to a height of 4 cm.
Samples were prepared and stored at both 4 and 25 °C. Samples
were analyzed using the Bruker Minispec Mq pulsed NMR unit
with the application sfc_lfc v2.51. Temperature of the NMR cham-
ber was controlled using a water bath pre set to desired
temperature.
2.8. X-ray diffraction
Static wide-angle X-ray diffraction of the emulsions was per-
formed on a Hecus S3-MICROcaliX (Hecus X-ray Systems GmbH,
Graz, Austria). The unit uses a GeniX 50 W microsource and cus-
tomized FOX-3D multi-layer point focussing optics (Xenocs SA,
Grenoble, France). Samples (20 ll) were loaded in quartz capillar-
ies (d = 1 mm) and characterized at 25 °C.
2.9. Contact angle measurements
A portion of the wax used to stabilize the emulsion was melted
at
80 °C with stirring, poured into an aluminium weigh boat at
room temperature and allowed to crystallize for
24 h. Solidified
ð1Þ wax discs were then cut into
1 cm3 cubes and placed in spec-
trophotometer cuvettes, with the smooth side facing upwards.
Cuvettes were then filled with light mineral oil alone or with
GMO. A small water droplet (D
1 mm) was injected from a nee-
dle (gauge 22, 0.394 mm nominal ID, Fisher Scientific, Nepean, ON,
Canada) onto the surface of the solid wax using a 10 ml syringe and
a programmable syringe pump (kd Scientific, Markham, ON,
Canada) operated at 0.1 ml/min. Images of sessile water droplets
were captured for 24 h using a Teli CCD camera with a macro lens
assembly and IDS Falcon/Eagle framegrabber (DataPhysics Instru-
ment GmbH, Filderstadt, Germany). Droplet contact angles against
the solid surface were measured digitally (SCA 20 software, version
2.1.5 build 16, DataPhysics Instrument GmbH, Filderstadt, Ger-
many). The oil phase/water interfacial tensions (IT) were deter-
mined using ASTM method D971 [18], with the DuNouy ring
method (Fisher surface tensiometer Model 21, Fisher Scientific,
Nepean, ON, Canada).
250 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
2.10. Statistical analyses
Triplicate analyses were performed on all sedimentation and
droplet size measurements. Analyses of variance and post hoc tests
were performed and statistical differences were deemed signifi-
cant at p = 0.05.
3. Results and discussion
3.1. Emulsion cooling rates
Fig. 3 shows a comparison of the different cooling rates used for
emulsions cooled to either 25 °C or 4 °C. A maximum cooling rate
of 4.6 °C/min was achieved with the quench-cooling and stirring
regime whereas the slowest cooling rate was with slow-static cool-
ing (0.7 °C/min). Stirring helped to maintain a more uniform cool-
ing rate throughout the beaker whereas statically-cooled samples
yielded slower temperature drops with some temperature inho-
mogeneities throughout. None of the cooling rates used were by
any means rapid. Thus, it would be incorrect to speak of true
‘quenching’ in any scenario.
3.2. Emulsion sedimentation
Control emulsions made with either GMO or wax destabilized
within 1 h of preparation. Hence, the presence of both the GMO
and paraffin wax was necessary to confer kinetic stability to all
emulsions. All wax-containing emulsions nevertheless demon-
strated phase separation dependent on composition and cooling/-
stirring regime (Figs. 4 and 5). When stored at 4 °C, the stirred
emulsions (quench or slowly-cooled) showed oil separation and
no aqueous phase separation. Separation was evident after 6 wks
(Fig. 4). Those slowly and statically-cooled broke down with a clear
water layer appearing soon after preparation. The quench-static
emulsions demonstrated limited phase separation (not readily vis-
ible in the beaker) over 78 wks. These results demonstrated that
more rapid formation of a wax crystal network prevented oil or
water separation. In regards to stirred emulsions, the encased dis-
persed phase appeared to sediment along with the wax crystal net-
work, leading to phase separation and oiling-off. The enhanced
sedimentation stability of the stirred emulsions with respect to
water separation ensured that the water droplets remained dis-
persed throughout the continuous oil phase during cooling and
wax crystallization. Stirring also appeared to promote more direct
contact between the aqueous droplets and wax crystals given the
Fig. 3. Temperature profiles of emulsions under the four cooling protocols
investigated. Cooling rates were calculated from the average slopes of the cooling
curves: quench-stirred (4.6 °C/min), quench-static (1.3 °C/min), slow-stirred
(0.9 °C/min), and slow-static (0.7 °C/min). Quench cooling was done in a 0 °C
waterbath whereas slow-cooled samples were placed in a 45 °C waterbath and
cooled at a rate of 1 °C/min. For cooling to 25 °C samples were withdrawn from the
waterbath when the centre of the sample reached desired temperature.
entrainment of the wax crystal network by the dispersed phase
in the stirred emulsions.
Emulsions cooled to 25 °C were less stable against sedimenta-
tion than at 4 °C (Fig. 5). All emulsions demonstrated some phase
separation after storage for 1.5 yrs, though the storage vessel used
(i.e., beaker vs. NMR tube) was critical to this evaluation. Stirred
emulsions (quench or slowly-cooled) showed oil separation after
storage for 1 wk, but no water phase separation even after
1.5 yrs. Quiescent emulsions (quench or slowly-cooled) demon-
strated some breakdown right after preparation, as the wax crystal
network formed was too sparse to encase the dispersed droplets.
With time, a supernatant oil phase was also visible. Storage in
NMR tubes accentuated the observed destabilization, which has
been ascribed to differences in wall effect and convective flow
[19]. The 25 °C protocol did not lead to measurable droplets with
NMR, further demonstrating the critical role that the wax crystal
network played on emulsion stabilization. Too weak a network
or too low an SWC resulted in an unstable emulsion, particularly
with so little emulsifier. Overall, there was a much slower rate of
emulsion sedimentation at 4 °C vs. 25 °C, as there was more solid
wax present to prevent density-driven settling of the dispersed
water droplets.
For comparison, the sedimentation stability of the emulsions’
oil phases was also evaluated separately. All samples remained
stable, and did not show any indication of phase separation at
4 °C and 25 °C. This lack of macroscopic phase separation clearly
indicated that the dispersed phase played a significant role on
phase separation. The continuous phase of all emulsions stored
at 25 °C was flowable whereas for those stored at 4 °C, only the
stirred samples were flowable, possibly due to the influence of stir-
ring on wax crystal size and inter-crystal interactions [20].
3.3. Sedimentation kinetics
Femulsion (Eq. (1)) plotted vs. storage time (Fig. 6) was used to
estimate sedimentation rate. Only stir-cooled emulsions were
evaluated as the statically-cooled emulsions rapidly destabilized
with clear separation of the aqueous phase. Amongst the stirred
emulsions, sedimentation was slowest for those stored at 4 °C
(no sedimentation for 1 wk) compared to storage at 25 °C (where
emulsions sedimented to
70% of their original heights within
the first wk). After 6 wks, the latter emulsions sedimented to

65% to 68% and to
63% to 65% after 1.5 years, respectively for
quenched and slowly-cooled systems. By contrast, the 4 °C emul-
sions sedimented to 82% and 78% of their initial heights after
6 wks for the quenched and slowly-cooled emulsions, respectively,
followed by a decrease in height to 69% for the quenched-stirred
emulsion and to 65% for the slow-stirred emulsion after 1.5 years.
3.4. Emulsion microstructure
Emulsion microstructure at 4 °C and 25 °C was highly-
dependent on the cooling/stirring regime used. As noted above,
emulsions stored at 4 °C were more stable against sedimentation
than those stored at 25 °C. Microscopy results concurred, with evi-
dence of a more stable dispersed aqueous phase at the lower stor-
age temperature. Fig. 7 shows the combined DIC and polarized
light micrographs of emulsions stored at 4 °C for 78 wks. Though
demonstrating oil separation, the stirred emulsions (quench and
slowly-cooled) showed the presence of a relatively stable dis-
persed aqueous phase, with little increase in droplet size over time.
By contrast, the statically-cooled emulsions underwent breakdown
with little evidence of dispersed droplets soon after emulsion
preparation. Irrespective of cooling rate, with stirring, there was
evidence of interfacial wax crystal adhesion (Fig. 7 arrows) sug-
gesting the existence of wax crystal surface activity, and the possi-
 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
Fig. 4. Sedimentation behaviour of emulsions stored at 4 °C as a function of time.
bility of GMO promoting interfacial wax crystallization. Such a
phenomenon was recently observed by us [11].
Fig. 8 shows combined DIC and polarized light micrographs of
emulsions stored at 25 °C for 78 wks. The quenched-stirred emul-
sion was the most stable against changes in droplet size, with a still
evident distribution of droplets visible after 1.5 years. The slow-
stirred emulsions were stable until at least 6 wks, but had broken
down after 78 wks. Both statically-cooled emulsions were not
stable, with the dispersed phase breaking down soon after homog-
enization and only a few water droplets can be seen. There was
some evidence of interfacially-adsorbed crystals in all cases, which
was most apparent with the quenched-stirred emulsions (as per
the 4 °C results). By comparing Figs. 7 and 8, it can be seen that ini-
tially, 4 °C emulsions consisted of smaller droplets compared with
those at 25 °C. After 6 wks, droplet sizes in emulsions at both tem-
peratures were comparable whereas after 1.5 years, only the 4 °C
emulsions remained stable with only a minor increase in droplet
size.
Fig. 9((A) and (B)) shows a close-up view of interfacial wax crys-
tallization promoted by stirring under quench-cooling conditions.
Clearly visible are interfacial crystals that stabilized the dispersed
phase via Pickering stabilization as well as network stabilization,
where crystals contributed to hindering droplet movement by
enmeshing them within a continuous crystalline matrix. Together,
these two mechanisms heightened the stability of the emulsions.
251
Macierzanka et al. [3] studied the time-dependent structural evo-
lution of W/O emulsions prepared with mixtures of two different
crystalline emulsifiers (partial acyglycerols and fatty acid mono-
diesters of propylene glycol). The emulsion droplets were initially
Pickering-stabilized. With time, interfacial crystals grew into the
continuous oil phase, leading to network stabilization and crys-
talline bridges between the droplets. Fig. 9A is the combined DIC
and polarized light image whereas Fig. 9B shows the corresponding
polarized light view only. Most crystalline species measured well
below 40 lm in length and existed as needles with high aspect
ratios. For comparison, Fig. 9C shows a putative crystal morphol-
ogy where the oil phase of the emulsion was quench-cooled to
4 °C under stirring. Other than the missing interfacial crystals
(given that no aqueous phase is present), the continuous phase
crystal appearance is quite similar to that observed in the
emulsions.
3.5. Evolution of droplet size distribution
Emulsion droplet sizes were only measurable for the stirred
emulsions cooled to 4 °C. All other emulsions (statically-cooled to
4 °C and all 25 °C emulsions) did not yield droplets measurable
via pulsed NMR, either due to their size (too large) or population
(too small). At 4 °C, the average droplet size (d33, 50th percentile
value) of the quench-stirred and slow-stirred emulsions ranged
252 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257

Fig. 5. Sedimentation behaviour of emulsions at 25 °C as a function of time. Statically-cooled emulsions were also stored in NMR tubes to show a clear view of aqueous phase
separation for both the quenched and slow-cooled samples (shown at right after 78 weeks’ storage).

free water content confirmed that both emulsions very slowly

broke down with time. The initial breadth of the quenched-
stirred emulsion droplet size distribution was broader than that
of its slowly-stirred counterpart. After 1.5 years, however, the lat-
ter’s droplet size distribution narrowed more extensively (i.e., 95%
of droplets were between 26 and 42 lm) which was accompanied
by an increase in % free water. This narrowing was interpreted as
greater emulsion destabilization than in the quenched-stirred
emulsion, which saw a smaller increase in free water during this
time.

3.6. Solid wax content

Fig. 11 shows the SWC of stirred emulsions as a function of

Fig. 6. Estimation of emulsion destabilization by sedimentation. Evolution of
percent emulsified layer (Femulsion from Eq. (1)) is plotted as a function of time for
slow or quenched-cooled stirred emulsions stored at 4 and 25 °C.
between 29 and 33 lm and remained unchanged over 78 wks
(Fig. 10) (p > 0.05). During this time, the 97.5th percentile values
significantly decreased indicating the loss of larger droplets as free
water (i.e., larger than the NMR instrumental limit of
500 lm) or
visually phase-separated water (p < 0.05). The gradual increase in
time. As the statically-cooled emulsions were not stable at either
4 °C or 25 °C, their SWC was not measured. The SWC was much
higher at 4 °C compared to 25 °C, with no effect of cooling rate
(p > 0.05). After 78 wks, there was a significant increase in SWC
at both temperatures due to continued wax solidification
(p < 0.05). Though crystallization to 4 °C led to initially higher
SWCs, over 78 wks, SWC increased to
60% for the emulsions pro-
cessed at 25 °C and only
20% for those at 4 °C.
Wax crystallization history was an important step in final SWC,
and it also contributed to the observed differences in wax crystal
 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
Cooling
Day 0
protocol
Quenched-
stirred
Quenched-
static
Slow-
stirred
Slow-
static
Fig. 7. Combined DIC and polarized light micrographs of wax-stabilized emulsions at 4 °C as a function of storage time. The presence of both interfacial crystals (shown by
arrows) and continuous phase crystals can be seen. The scale bar represents 40 lm.
morphology and population. Irrespective of the stirring/cooling
regime, crystallization began when supercooling was sufficient
for the highest-melting wax species to nucleate and act as crystal-
lization centres. With more nuclei formed, there was more rapid
bulk crystallization. Given the compositional variety of the wax
(Fig. 1), low supercooling (i.e., cooling to 25 °C) initiated high-
melting wax crystallization first. High supercooling (4 °C) likely
resulted in more rapid, mass crystallization of the bulk, with for-
mation of more mixed crystals consisting of lower and higher-
melting species, with perhaps these crystals acting as secondary
nucleation sites.
3.7. Wax crystal structure
X-ray diffraction revealed that the lamellar spacing of the solid
paraffin wax at 25 °C was 41.57 Å with a weak second-order spac-
ing at 20.39 Å and short spacings at 4.18 Å and 3.74 Å. Little change
in both small and long-order spacings were observed for the wax
present in all emulsions, with first order spacing at
42.1 to
42.5 Å. Short spacings were identical to the solid wax, with spacings
at 4.18 Å and 3.74 Å. Given the alkane diversity present in the paraf-
fin wax (straight-chained, branched, etc.), it would be imprudent to
assign specific crystalline structures to each of these samples. Based
on Craig et al. for pure alkanes, the 100% paraffin wax existed in the
orthorhombic unit cell [21]. The orthorhombic unit cell c/2 value,
which corresponds to the dimension of the carbon atom chains
253
Week 1 Week 6 Week 78
and gaps between the planes of methyl endgroups between two
successive molecular layers was determined using:
ðc=ÅÞ=2 ¼ 1:2724nc ¼ 1:8752 ð2Þ
where nc is the average chain length. Based on Eq. (2), a (c/Å)/2
spacing of 41.57 Å indicates an average alkane length (nc) of
31
carbons, which closely matched the compositional analysis of wax
with an average alkane chain length of 29 carbons (Fig. 1).
3.8. Wax crystal wetting
Contact angles (h) between the solid wax, and aqueous and oily
phases were used to understand the ability of wax to promote
Pickering crystallization. Equilibrium h of water droplets on the
wax surface were measured at 147.5 ± 2.9° in light mineral oil
and 130.5 ± 4.1° with the added presence of GMO. It is well-
known that 90° < h < 180° favours stabilization of water droplets
in an oil-continuous medium [2]. At h = 90°, particles are equally
wetted by the water and oil phases giving rise to the most stable
emulsion. At h = 180°, crystals are completely wetted by the oil
phase so no association with the water droplets occurs. In the pre-
sent case, the smallest h measured through the aqueous phase was
obtained with GMO-doped mineral oil against wax. The IT of min-
eral oil against water was 44.9 ± 3.4 mN/m and 4.1 ± 0.3 with
added presence of GMO.
254 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
Cooling
Day 0 Week 1
protocol
Quenched-
stirred
Quenched-
static
Slow-
stirred
Slow-
static
Fig. 8. Combined DIC and polarized light micrographs of wax-stabilized emulsions at 25 °C as a function of storage time. The presence of both interfacial crystals (shown by
arrows) and continuous phase crystals can be seen. The scale bar represents 40 lm.
The displacement energy (Edisp), which is directly linked to h,
was used as a measure of the energy required to remove wax crys-
tals from the oil/water interface [22] (Eq. (3)):
Edisp ¼ pr 2 cow ð1  cos hÞ2
where cow is the IT at the oil–water interface and r is the radius of a
wax ‘particle’. The sign inside the bracket is negative for displace-
ment into the water phase, and positive for displacement into oil
(e.g., movement of wax crystals into the oil phase) [23]. From Eq.
(3), Edisp increases with an increase in oil–water IT, but decreases
with increase in h. Assuming r
1 lm, Edisp to remove a wax crystal
from the water/GMO/mineral oil interface was 3.8  108 kT, which
indicates that presence of wax crystals at the water droplet surface
is favourable unless a high amount of energy is provided for
removal. Of note, Edisp to remove a wax crystal from the water/min-
eral oil interface (without GMO) was 8.5  108 kT, comparable to
the value for GMO. However, as molten wax has no emulsifying
activity, it could not be used to form and stabilize water droplets,
hence, their movement towards the interface is not favourable.
3.9. Wax crystallization and network structure
Effective emulsion stabilization was dependent on the con-
certed roles of interfacial and network wax crystallization, with
both contributing to delaying water droplet sedimentation, floccu-
lation and coalescence. Such ‘dual-mode’ wax crystallization was
Week 6 Week 78
itself dependent on temperature and mixing regime during emul-
sion processing. Crystallization to 4 °C yielded SWCs 3–4  higher
than at 25 °C, which markedly impacted emulsion stability. At
lower SWCs, there was insufficient solid mass to effectively sur-
ð3Þ round the droplets and physically entrap the aqueous phase
whereas the higher SWC at 4 °C retarded phase separation and coa-
lescence. The presence of both interfacial and surrounding crystals
kept droplets separate from one another due to the presence of
interstitial crystals thus reducing the rate of flocculation. However,
wax that crystallized directly at the interface visibly retarded coa-
lescence and sedimentation vis-à-vis network-type crystals, as
these crystals enveloped individual droplets thus preventing film
drainage and rupture. As wax is not surface-active, the use of the
surfactant GMO promoted direct interfacial crystallization. By con-
trast, bulk network crystallization, though also beneficial in stabi-
lizing the emulsions, did not offer the same level of stabilization
against emulsion breakdown, presumably due to gravity-induced
droplet movement throughout the wax crystal network.
Pickering particle removal from an oil–water interface is a nec-
essary precursor to coalescence. The energy required for crystal
displacement will depend on the composition and rheology of
the interface as well as crystal wettability [24]. There was no par-
ticle desorption from the interface observed even after 78 wks
given the high Edisp. Rather, alkane self-organization and crystal-
lization promoted the formation of a solid shell that englobed
the dispersed droplets. This suggested that incomplete coverage
 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257 255

Fig. 9. Combined DIC and polarized light micrograph (A) and corresponding polarized light only micrograph (B) of wax-stabilized emulsions at 4 °C. The bottom picture
presents a polarized light micrograph of a wax–oil mixture (C) at 4 °C cooled using a similar protocol to the emulsions. Scale bar represents 40 lm.

Fig. 10. Droplet size distribution and free water content of quenched-stirred and slow-stirred emulsions at 4 °C as a function of time. 2.5th, 50th, and 97.5th percentile
droplet size are shown from left to right on day 0 and week 78.

and/or absence of interfacial crystallization in some emulsions was
responsible for film drainage and coalescence [25], particularly if
little affinity between the droplet surface and wax crystals existed,
as observed in the quiescently-cooled samples. Thus, mixing
regime was critical to effective emulsion stabilization. In the
absence of mixing during cooling, little or no interfacial crystalliza-
tion was observed, given the lack of wax mass transfer towards the
GMO-covered droplet surface.
As wax is not surface-active, the GMO, other than its role in
reducing interfacial tension, promoted wax interfacial adhesion.
This behaviour is similar to recent reports showing the concerted
role of surfactants and particles in Pickering-type stabilization,
256 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
Fig. 11. Solid wax content of quench-stirred and slow-stirred emulsions at 25 °C
and 4 °C as a function of time.
where interaction between the former and latter leads to enhanced
emulsion stabilization [26]. However, this primarily occurred with
concurrent cooling and mixing, where interaction between the
oleic acid in the GMO and aliphatic chains in the wax was pro-
moted. Such chain–chain ordering encouraged heterogeneous
nucleation of the wax at the interface, as noted by Davies et al.
in their work on GMO-tristearin systems [27]. Similarly, Arima
et al. found that high-melting sucrose palmitic acid oligoesters
acted as a template for the interfacial nucleation and growth of
palm oil mid-fraction at the surface of oil droplets in oil-in-water
emulsions [28]. These now-anchored wax crystals provided a
physical barrier against flocculation and coalescence. The GMO
promoted interfacial crystallization either via modification of the
wax’s wetting behaviour or by promoting heterogeneous nucle-
ation. Other efforts in our laboratory have demonstrated that sur-
factant structural complementarity is an important parameter in
whether species such as wax or triacylglycerols will interfacially
crystallize or not [17].
4. Conclusion
Though the scientific literature is replete with research efforts
on Pickering emulsions stabilized with wax (e.g., Li et al. [29]), little
is known about the concerted role of wax crystals simultaneously
present at the oil–water interface and within the continuous oil
phase. The present research clearly demonstrates that control of
the interactions between dispersed crystals and neighbouring
aqueous droplets in W/O emulsions may be used as a tool to mod-
ulate the long-term stability of W/O emulsions. Such findings are
significant given that oil-continuous emulsions are commonplace
in cosmetics, personal care products, crude oil and processed foods
[30].
General requirements dictating successful emulsion stabiliza-
tion by Pickering particles include optimization of their spatial
arrangement, size and surface activity as well as the extent of dro-
plet coverage [31]. The wax crystals in this study englobed the dis-
persed aqueous droplets with a thin mechanical barrier. They also
performed ‘double-duty’ by embedding themselves within the
interface of two or more neighbouring droplets thereby arresting
coalescence via either monolayer or multilayer bridging. Wax crys-
tal adhesion to the droplet surface required the presence of a sur-
factant, in this present case GMO, to confer surface activity to the
wax. Beyond the ability of the interfacially-adsorbed crystals to
hinder interdroplet coalescence, wax crystals with little to no affin-
ity for the oil–water interface aggregated to form a network in the
continuous oil phase, which reduced water droplet settling due to
gravity differences. Finally, to permit effective coverage, the diam-
eter of interfacially-adsorbed particles should be much smaller
than the droplet size to permit effective enrobing which, generally
speaking, was the case with the wax crystals.
The present results showed that the long-term stability of the
present wax-stabilized W/O emulsions was highly-dependent on
the concerted role of interfacial and network wax crystallization,
with the combination of stirring and quench-cooling providing
the highest kinetic stability. Further elucidation of the mechanisms
at play is necessary to clarify the behaviour of wax species at the
oil–water interface, in particular the possible role of surfactant-
wax molecular complementarity as well as the physical state of
the interfacially-bound surfactant. Though solid-state templating
is clearly established as a mechanism for heterogeneous nucle-
ation, these results suggest that liquid-state surfactants may also
promote such templating. This is an on-going challenge currently
being addressed.
Acknowledgment
Financial assistance from the Natural Sciences and Engineering
Research Council of Canada (NSERC) is gratefully acknowledged.
References
[1] A. Beri, J.E. Norton, I.T. Norton, Effect of emulsifier type and concentration,
aqueous phase volume and wax ratio on physical, material and mechanical
properties of water in oil lipsticks, Int. J. Cosmet. Sci. 35 (6) (2013) 613–621.
[2] B. Binks, Particles as surfactants—similarities and differences, Curr. Opin.
Colloid Interface Sci. 7 (2002) 21–41.
[3] A. Macierzanka, H. Szela˛g, P. Szumała, R. Pawłowicz, A.R. Mackie, M.J. Ridout,
Effect of crystalline emulsifier composition on structural transformations of
water-in-oil emulsions: emulsification and quiescent conditions, Colloids Surf.,
A 334 (1–3) (2009) 40–52.
[4] D. Rousseau, S. Ghosh, H. Park, Comparison of the dispersed phase coalescence
mechanisms in different tablespreads, J. Food Sci. 74 (1) (2009) E1–E7.
[5] G. Sun, J. Zhang, H. Li, Structural behaviors of waxy crude oil emulsion gels,
Energy Fuels 28 (6) (2014) 3718–3729.
[6] D. Tambe, M. Sharma, The effect of colloidal particles on fluid–fluid interfacial
properties and emulsion stability, Adv. Colloid Interface Sci. 52 (1994).
[7] M.T. Ghannam, Water-in-crude oil emulsion stability investigation, Pet. Sci.
Technol. 23 (January) (2005) 649–667.
[8] D. Johansson, B. Bergenståhl, E. Lundgren, Wetting of fat crystals by
triglyceride oil and water. 1. The effect of additives, J. Am. Oil Chem. Soc. 72
(8) (1995) 921–931.
[9] S.M. Hodge, D. Rousseau, Continuous-phase fat crystals strongly influence
water-in-oil emulsion stability, J. Am. Oil Chem. Soc. 82 (3) (2005) 159–164.
[10] E.H. Lucassen-Reynders, M. van den Tempel, Stabilization of water-in-oil
emulsions by solid particles, J. Phys. Chem. 67 (4) (1963) 731–734.
[11] S. Ghosh, D. Rousseau, Triacylglycerol interfacial crystallization and shear
structuring in water-in-oil emulsions, Cryst. Growth Des. 12 (10) (2012) 4944–
4954.
[12] S. Haj-shafiei, S. Ghosh, D. Rousseau, Kinetic stability and rheology of wax-
stabilized water-in-oil emulsions at different water cuts, J. Colloid Interface
Sci. 410 (2013) 11–20.
[13] R.F.G. Visintin, T.P. Lockhart, R. Lapasin, P. D’Antona, Structure of waxy crude
oil emulsion gels, J. Nonnewton. Fluid Mech. 149 (1–3) (2008) 34–39.
[14] S.M. Hodge, D. Rousseau, Flocculation and coalescence in water-in-oil
emulsions stabilized by paraffin wax crystals, Food Res. Int. 36 (7) (2003)
695–702.
[15] D. Rousseau, S.M. Hodge, Stabilization of water-in-oil emulsions with
continuous phase crystals, Colloids Surf., A 260 (1–3) (2005) 229–237.
[16] M. Pradhan, D. Rousseau, A one-step process for oil-in-water-in-oil double
emulsion formation using a single surfactant, J. Colloid Interface Sci. 386 (1)
(2012) 398–404.
[17] S. Ghosh, T. Tran, D. Rousseau, Comparison of Pickering and network
stabilization in water-in-oil emulsions, Langmuir 27 (11) (2011) 6589–6597.
[18] ASTM D971-99a Standard Test Method for Interfacial Tension of Oil Against
Water by the Ring Method; see <http://www.astm.org/Standards/D971.htm>
(accessed 01.04.15).
[19] P. Mazur, W. van Saarloos, Many-sphere hydrodynamic interactions and
mobilities in a suspension, Physica A 82 (1982). 90127-3.
 S. Ghosh et al. / Journal of Colloid and Interface Science 460 (2015) 247–257
[20] Y. Zhang, J. Gong, Y. Ren, P. Wang, Effect of emulsion characteristics on wax
deposition from water-in-waxy crude oil emulsions under static cooling
conditions, Energy Fuels 24 (2) (2010) 1146–1155.
[21] S.R. Craig, G.P. Hastie, K.J. Roberts, A.R. Gerson, J.N. Sherwood, R.D. Tack,
Investigation into the structures of binary-, tertiary- and quinternary-mixtures
of n-alkanes and real diesel waxes using high-resolution synchrotron X-ray
powder diffraction, J. Mater. Chem. 8 (1998) 859–869.
[22] B.P. Binks, S.O. Lumsdon, Influence of particle wettability on the type and
stability of surfactant-free emulsions, Langmuir 16 (23) (2000)
8622–8631.
[23] T. Tran, S. Ghosh, D. Rousseau, Spheroidal fat crystal microstructures formed
with confined gap shearing, Cryst. Growth Des. 14 (12) (2014) 6383–6390.
[24] D. Rousseau, Fat crystals and emulsion stability—a review, Food Res. Int. 33
(2000) 3–14.
[25] J. Frelichowska, M.-A. Bolzinger, Y. Chevalier, Effects of solid particle content
on properties of o/w Pickering emulsions, J. Colloid Interface Sci. 351 (2)
(2010) 348–356.
257
[26] A. Nesterenko, A. Drelich, H. Lu, D. Clausse, I. Pezron, Influence of a mixed
particle/surfactant emulsifier system on water-in-oil emulsion stability,
Colloids Surf., A 457 (2014) 49–57.
[27] E. Davies, E. Dickinson, R. Bee, Shear stability of sodium caseinate emulsions
containing monoglyceride and triglyceride crystals, Food Hydrocolloids 14
(2000) 145–153.
[28] S. Arima, S. Ueno, a. Ogawa, K. Sato, Scanning microbeam small-angle X-ray
diffraction study of interfacial heterogeneous crystallization of fat crystals in
oil-in-water emulsion droplets, Langmuir: ACS J. Surf. Colloids 25 (17) (2009)
9777–9784.
[29] C. Li, Z. Mei, Q. Liu, J. Wang, J. Xu, D. Sun, Formation and properties of paraffin
wax submicron emulsions prepared by the emulsion inversion point method,
Colloids Surf., A 356 (1–3) (2010) 71–77.
[30] B.P. Binks, R. Murakami, Phase inversion of particle-stabilized materials from
foams to dry water, Nat. Mater. 5 (11) (2006) 865–869.
[31] E. Dickinson, Use of nanoparticles and microparticles in the formation and
stabilization of food emulsions, Trends Food Sci. Technol. 24 (1) (2012) 4–12.