1130 Langmuir 2004, 20, 1130-1137

Silica Particle-Stabilized Emulsions of Silicone Oil and

Water: Aspects of Emulsification
Bernard P. Binks* and Catherine P. Whitby
Surfactant & Colloid Group, Department of Chemistry, University of Hull,
Hull, HU6 7RX, United Kingdom

Received September 10, 2003. In Final Form: November 20, 2003

A study of the emulsification of silicone oil and water in the presence of partially hydrophobic, monodisperse
silica nanoparticles is described. Emulsification involves the fragmentation of bulk liquids and the resulting
large drops and the coalescence of some of those drops. The influence of particle concentration, oil/water
ratio, and emulsification time on the relative extents of fragmentation and coalescence during the formation
of emulsions, prepared using either batch or continuous methods, has been investigated. For batch emulsions,
the average drop diameter decreases with increasing particle concentration as the extent of limited
coalescence is reduced. Increasing the oil volume fraction in the emulsion at fixed aqueous particle
concentration results in an increase in the average drop diameter together with a dramatic lowering of
the uniformity of the drop size distribution as coalescence becomes increasingly significant until catastrophic
phase inversion occurs. For low oil volume fractions (φo), fragmentation dominates during emulsification
since the mean drop size decreases with emulsification time. For higher φo close to conditions of phase
inversion, coalescence becomes more prevalent and the drop size increases with time with stable multiple
emulsions forming as a result.

Introduction presently unclear how both types of drops can be stabilized

Macroemulsions are mixtures of two immiscible liquids,
in which one is dispersed as micrometer-sized drops within
the continuous phase of the other. Nanoparticles are one
type of additive which can be used to populate the liquid-
liquid interface formed during homogenization.1,2 Al-
though the interfacial tension is probably unaffected by
particle adsorption (the surface tension of colloidal
suspensions is typically close to that of water3), the area
of the bare oil-water interface is reduced, which lowers
the amount of energy required to form the interface. The
free-energy change associated with the transfer of an
isolated, spherical particle from a bulk water or oil phase
to a planar oil-water interface is at a maximum for a
particle which is wet to an intermediate extent by both
liquid phases.4,5 Such particles are effectively irreversibly
adsorbed at the liquid-liquid interface and, if the drops
are covered sufficiently, are responsible for sterically
hindering drop coalescence, thus stabilizing the emulsion.
The type of drops formed (oil or water) is primarily
determined by the wettability of the particles. It is thought
that, in an analogous fashion to the behavior of surfactant
layers at liquid-liquid interfaces, the particle layer curves
such that the larger fraction of the particle surfaces resides
on the external surface of the drop in the liquid which
preferentially wets the particles.1 It is possible, however,
to invert the emulsion type by, for example, gradually
increasing the volume fraction of the dispersed phase,6
causing a so-called catastrophic phase inversion.7 It is
by the same type of particles.
Emulsion formation can be seen simplistically as
consisting of two processes: fragmentation of the bulk
liquids and the resulting large drops and the possible
coalescence of such drops. In the solid-stabilized emulsions,
the extent of fragmentation seems to be determined by
the amount of energy supplied during emulsification.
Typically, it is assumed that the fragmented drops coalesce
together until the particle coverage at the drop interfaces
is sufficient to prevent further coalescence.8 Recently, it
was proposed that phase inversion in liquid-liquid
dispersions occurs via the formation of multiple globules
as the dispersed phase volume fraction is increased.9 These
form as drops of the continuous phase are incorporated
into drops of the dispersed phase during coalescence. The
formation of multiple emulsions was also shown to lead
to inversion of the continuous phase where the emulsion
is homogenized for sufficiently long times.10 These ideas
have a clear application in the understanding of phase
inversion in surfactant-stabilized and solid-stabilized
emulsions. A variety of factors affect the extent of
coalescence during emulsification of solid-stabilized emul-
sions, including the particle concentration, time of emul-
sification, and the dispersed phase volume fraction. We
report here a series of experiments investigating the effects
of these parameters on the formation of solid-stabilized
emulsions of silicone oil and water and, in particular, on
the balance achieved between the processes of fragmen-
tation and coalescence during homogenization.

* To whom correspondence should be addressed. E-mail: Experimental Section

b.p.binks@hull.ac.uk.
(1) Binks, B. P. Curr. Opin. Colloid Interface Sci. 2002, 7, 21.
(2) Aveyard, A.; Binks, B. P.; Clint, J. H. Adv. Colloid Interface Sci.
2003, 100-102, 503.
(3) Okubu, T. J. Colloid Interface Sci. 1995, 171, 55.
(4) Levine, S.; Bowen, B. D.; Partridge, S. J. Colloids Surf. 1989, 38,
325.
(5) Binks, B. P.; Lumsdon, S. O. Langmuir 2000, 16, 8622.
(6) Binks, B. P.; Lumsdon, S. O. Langmuir 2000, 16, 2539.
(7) Salager, J.-L. In Encyclopedia of Emulsion Technology; Becher,
P., Ed.; Marcel Dekker: New York, 1988; Vol. 3, p 79.
(i) Materials. Water was first passed through an Elga reverse
osmosis unit and then a Milli-Q reagent water system. The oils
used were 350 cS (density ) 0.97 g cm-3 at 25 °C) and 0.65 cS
(8) Whitesides, T. H.; Ross, D. S. J. Colloid Interface Sci. 1995, 169,
48.
(9) Pacek, A. W.; Nienow, A. W.; Moore, I. P. T. Chem. Eng. Sci. 1994,
49, 3485.
(10) Groenweg, F.; Agterof, W. G. M.; Jaeger, P.; Janssen, J. J. M.;
Wiering, J. A.; Klahn, J. K. Trans. Inst. Chem. Eng. 1998, 76, 55.

10.1021/la0303557 CCC: $27.50 © 2004 American Chemical Society

Published on Web 01/22/2004
Silica Particle-Stabilized Emulsions Langmuir, Vol. 20, No. 4, 2004 1131

(density ) 0.76 g cm-3) poly(dimethylsiloxane) (PDMS, Aldrich),

of purity greater than 99%. The less viscous oil was columned
twice through chromatographic alumina to remove polar impuri-
ties. The partially hydrophobic silica particles were a gift from
Rhodia Services, with a primary particle diameter of 25 nm.
They were prepared by grafting n-octyltriethoxysilane onto
hydrophilic silica (purchased from Clariant). The particles were
transferred from the synthesis media into water by dialysis. After
dialysis, the particle concentration in the dispersion was 0.7 wt
%. The background electrolyte, sodium chloride (Prolabo), was
used as received (analytical grade).
(ii) Methods. The emulsions were prepared using a Janke
and Kunkel Ultra-Turrax homogenizer with an 18 mm head
operating at 13 500 rpm. For experiments investigating the effect
of particle concentration, all emulsions were prepared by
homogenizing the water and oil phases (at an oil volume fraction
of 0.33) together for 3 min. For experiments investigating the
effect of oil volume fraction, the emulsions were prepared in one
of two ways. In some experiments, individual emulsions were
prepared for each oil volume fraction studied by homogenizing
oil and water together for 3 min (batch). In others, the oil volume
fraction of the emulsion was increased sequentially. In this
continuous method of preparation, an emulsion with φo ) 0.1
was prepared by homogenization for 3 min. The oil volume fraction
of the emulsion was then increased in steps of 0.1, re-
homogenizing for 1 min after each addition of oil.
The emulsion type was inferred by observing whether a drop Figure 1. TEM image of the monodisperse, spherical, partially
of the emulsion dispersed when added to a small volume of water hydrophobic silica nanoparticles. The bar represents 100 nm.
or oil. The final emulsions obtained from all the experiments
were transferred into stoppered, graduated vessels and ther-
mostated at 25 °C. The stability over time of some of the emulsions transferred into a low-temperature chamber (Oxford Instru-
prepared was determined by monitoring the appearance and ments CP2000) maintained under ultrahigh vacuum where the
movement of the oil-emulsion and emulsion-water interfaces. emulsion was fractured at -95 °C and etched for 3 min. The
The emulsion drop size distributions were examined both by exposed emulsion surface was coated with a thin layer of a mix-
light scattering and optical microscopy techniques. The volume ture of gold and palladium. The stub was then transferred into
weighted droplet diameter distribution of the emulsion was an FE-SEM (JEOL 6301F) fitted with a cold stage (Cressing-
determined using a Malvern Mastersizer 2000. The emulsion ton Instruments) operating at about -150 °C. Images were
was diluted in either oil or water, as appropriate. We are confident captured using Analysis and Soft Imaging ADDA II system
that dilution does not change the drop size distribution since software.
particles are effectively irreversibly adsorbed.6 In the water
continuous emulsions, the light scattering measurements were Results and Discussion
done with a dilute sample being cycled continuously through the A variety of particles of different size, shape, and surface
optical unit, while dilute oil continuous emulsions were simply chemistry have been used to stabilize emulsions including
injected into the optical unit to minimize the volume of oil re-
hydrophobic silicas, clays, carbon nanotubes, and latexes.
quired for the measurement. The emulsions were characterized
by the mean drop diameter and the uniformity, which is a measure Considerable work has already been done with emulsions
of the polydispersity, of the distribution. The uniformity, U, is stabilized by partially hydrophobic silica particles;1 thus,
given by the characteristics and stability of such emulsions are
well known. These studies, however, employed an amor-

∑N d 3 phous fumed silica powder consisting of polydisperse

|dm - di|
i i
1 i
U) (1)
dm
∑N d i i
3
i TEM image shown in Figure 1. In water the particles
where Ni is the number of drops with diameter di and dm is the
median drop diameter of the distribution. A small, dilute sample
of each emulsion was also added to a haemocytometer cell (Weber
Scientific) and viewed with a Nikon Labophot microscope fitted
with a DIC-U (World Precision Instruments) camera. The images
were processed using Adobe Photoshop 5.0 software.
Transmission electron microscopy (TEM) of the silica nano-
particles was performed using a JEOL 100C 80 kV electron
microscope. A single drop of a dilute aqueous dispersion of the
particles was placed on a carbon-coated copper grid (300 mesh)
and allowed to evaporate.
The structure of the adsorbed particle layer at the oil-water
interface of some of the emulsions was imaged using low-
temperature field emission scanning electron microscopy
(LTFESEM) following the methodology previously described.11
A drop of emulsion was mounted in an aluminum low-temper-
ature SEM stub and frozen in nitrogen slush. The stub was then
(11) Binks, B. P.; Kirkland, M. Phys. Chem. Chem. Phys. 2002, 4,
3727.
particles of quasi-spherical shape which may fuse together
at close separations. The partially hydrophobic silica
nanoparticles used here are monodisperse and spherical
of about 30 nm in diameter, which is confirmed by the
flocculate strongly, forming aggregates with a wide range
of sizes which sediment rapidly. Adding salt (up to 1 M)
to the dispersion or adjusting its pH between 3 and 9 does
not affect the rate or extent of the sedimentation.
Increasing the particle concentration in the dispersion up
to 5 wt % increases the viscosity of the dispersion slightly,
slowing down sedimentation.
These particles readily stabilize oil-in-water (o/w)
emulsions implying that the extent of hydrophobization
is not high. Emulsions of 33 vol % 350 cS PDMS in 10-2
M NaCl stabilized by 0.7 wt % of the hydrophobic silica
in the aqueous phase release about 60% of the water phase
in 4 h (because of creaming) but show no evidence of
coalescence even after long times (6 months). The stability
of the emulsions and the average drop diameter (about 50
µm) do not change substantially as the pH (3-9) or the
background electrolyte concentration (10-3 M to 1 M) of
the aqueous phase is varied. This is consistent with the
insensitivity of the stability of the aqueous particle
1132 Langmuir, Vol. 20, No. 4, 2004
Figure 2. Effect of particle concentration on the fraction of
water released by 33 vol % 350 cS PDMS-in-10-2 M NaCl
emulsions stabilized by hydrophobic silica because of creaming
as a function of time since preparation. The concentration of
particles in the aqueous phases of the emulsions are (from top
to bottom) 0.5, 1, 2, and 5 wt %.
dispersion to changes in the electrolyte concentration
and pH.12
(i) Effect of Particle Concentration. The effect of
particle concentration on the stability of 33 vol % 350 cS
PDMS in 10-2 M NaCl emulsions stabilized by hydro-
phobic silica in the aqueous phase at natural pH (∼6) is
shown in Figure 2. The ordinate is the fraction of the
aqueous phase which is resolved because of creaming. At
low particle concentrations, just over half of the water
is resolved from the emulsion over about 6 h. As the par-
ticle concentration in the emulsions is increased, the ex-
tent of creaming decreases and the time taken for the
emulsion to cream increases until at 5 wt % only a small
fraction (<0.1) of water is released over about 8 h. The
rate at which an emulsion drop creams depends both on
the drop diameter and the viscosity of the continuous
medium. The increase in stability of the emulsions to
creaming is due, at low particle concentrations, to the
decrease in average drop diameter (shown later). At higher
particle concentrations, where the average drop diameter
is constant (also shown later), it is likely to be due to the
slight increase in viscosity of the continuous aqueous
phase.
Figure 3 shows the drop size distribution of the
emulsions for different particle concentrations in water.
At low particle concentrations, we observe a log-normal
drop size distribution corresponding to a single population
of drops centered at about 50 µm. Clearly, the emulsifica-
tion process is already efficient at producing a relatively
uniform distribution of drop sizes. As the particle con-
centration increases, the distribution shifts to smaller
diameters. At high particle concentrations, the emulsion
drop size distributions are increasingly asymmetric and
show evidence of a shoulder at smaller diameters (e10
µm). Consequently, the average drop diameter of the
emulsion decreases as a function of increasing particle
concentration.
It has been found previously that the average drop size
of solid-stabilized emulsions decreases with increasing
particle concentration as more particles are available to
(12) Binks, B. P.; Lumsdon, S. O. Phys. Chem. Chem. Phys. 1999, 1,
3007.
Binks and Whitby
Figure 3. Drop size distributions for 33 vol % 350 cS PDMS-
in-10-2 M NaCl emulsions stabilized by 0.5 wt % (full line), 1.5
wt % (dot-dashed line), 2 wt % (dotted line), and 5 wt % (dashed
line) hydrophobic silica in the aqueous phase.
stabilize smaller drops.6,13,14 It seems reasonable to assume
that, for given homogenization conditions, the phase being
broken down into drops is fragmented initially to the same
extent, irrespective of the particle concentration. Depend-
ing on the amount of particles adsorbed on the drop
surfaces, the drops will coalesce together, reducing the
total drop surface area, until the particle coverage is
sufficient to stabilize the drops against further coalescence.
On the basis of the drop size distribution and assuming
the adsorbed particles form a hexagonal close-packed layer
at the drop surface, the number of particles (na) required
to form a single layer completely covering the available
drop surfaces can be calculated and compared to the total
number of particles available (nt). In Figure 4a, both the
variation in the average emulsion drop diameter and the
ratio nt/na are shown as a function of particle concentration.
As discussed previously, the average drop diameter
decreases as the particle concentration increases, with
the drop diameter becoming almost constant at concen-
trations above approximately 3.5 wt %. The ratio of
particles available in the emulsion to that required to
cover all the drops with a monolayer rises dramatically
as the particle concentration increases. Particles in excess
of that required for monolayer coverage may adsorb to
form further layers or remain dispersed in the continuous
phase (see later).
This behavior can be examined further in the light of
the effect of particle concentration on emulsions of a less
viscous oil, 0.65 cS PDMS, in water. These emulsions were
stabilized by silica particles which had been silanized to
a slightly lesser extent, but which were for the purposes
of this study the same as those used above. In Figure 4b,
the variations in the average emulsion drop diameter and
the ratio nt/na are shown as a function of particle
concentration. For this system, as the particle concentra-
tion increases up to 3 wt %, the drop diameter decreases
by a factor of about 5 while the ratio of particles available
to those required for monolayer coverage of the drops
remains at around unity. In this concentration regime,
the drop size is limited by the particle concentration. As
the particle concentration in the emulsion is increased
further up to 5.6 wt %, the drop diameter remains
approximately constant while the ratio increases to ≈2.
Thus, there is an excess of particles, some of which may
(13) Moore, W. C. J. Am. Chem. Soc. 1919, 41, 940.
(14) Midmore, B. R. J. Colloid Interface Sci. 1999, 213, 352.
Silica Particle-Stabilized Emulsions
Figure 4. The median drop diameter (left-hand ordinate,
circles) of (a) 33 vol % 350 cS PDMS-in-10-2 M NaCl emulsions
and (b) 33 vol % 0.65 cS PDMS-in-water emulsions stabilized
by hydrophobic silica as a function of the concentration of
particles in the aqueous phase. Also shown is the ratio of the
total number of particles available, nt, to the number required
to provide a monolayer around all drops, na (right-hand ordinate,
diamonds).
reside in the continuous phase, which was confirmed by
the presence of a small shoulder centered at about 1 µm
in the drop size distributions, attributed to aggregated
silica particles.
It is unclear why the trends in the ratio nt/na with
particle concentration are so different for these two
emulsions. At high concentrations, an excess of particles
in the continuous phase is likely. The larger excess in
emulsions of the more viscous silicone oil is presumably
due in part to the high viscosity of the oil hindering its
fragmentation into smaller drops which could also be
stabilized. At low particle concentrations, where it is
thought that the drop size is set by the particle concen-
tration, a roughly constant particle excess would be
expected and is observed for emulsions of the less viscous
oil. As will be shown later, drops of 350 cS PDMS tend not
to relax back into a spherical shape after distortion by the
action of the rotor-stator homogenizer. Should this distor-
tion of the drops result in an inhomogeneous distribution
of particles at the drop surfaces, this might favor drop
coalescence to a greater extent during homogenization.
Langmuir, Vol. 20, No. 4, 2004 1133
Figure 5. LTFESEM images of 33 vol % 350 cS PDMS-in-10-2
M NaCl emulsions stabilized by 0.5 wt % (upper), 1.5 wt %
(middle), and 5 wt % (lower) hydrophobic silica particles in the
aqueous phase. All images are focused on the surface of an
emulsion drop and show the layers of particles adsorbed at the
oil-water interface. In the upper image, a small segment of a
drop surface revealed by the cleaving of the sample is shown.
In the middle image, the upper surface of a drop can be seen.
In the lower image, the side of a drop to which a few smaller
drops are clustered is shown. Notice in the bottom right-hand
corner, excess silica particles not attached to the oil-water
interface can be seen. The bars represent 500 nm.
The assumption that the drops require a close-packed
layer of particles to be stabilized is supported by observing
the continuous, densely packed layers of particles at the
oil-water interface in emulsion samples studied by
LTFESEM, examples of which are shown in Figure 5. The
continuous phase of the emulsions has been etched to
reveal the oil drops and, in particular, the surface of the
drops. At all particle concentrations, the adsorbed layer
consists of spherical particles approximately 30 nm in
1134 Langmuir, Vol. 20, No. 4, 2004
Figure 6. Variation of the mean drop diameter (left-hand
ordinate, squares) and uniformity (right-hand ordinate, tri-
angles) of 350 cS PDMS-10-2 M NaCl emulsions stabilized by
0.7 wt % hydrophobic silica in the aqueous phase with oil volume
fraction (varied continuously). The emulsion underwent cata-
strophic phase inversion from a water continuous (open points)
to an oil continuous (filled points) one at an oil volume fraction
of 0.67 (dotted line).
diameter, indicating that the large flocs of particles in the
original aqueous dispersion are broken down into the
primary particles during emulsification. The layer is
increasingly densely packed as the particle concentration
in the emulsion is raised. It seems that the drops require
a dense layer of particles to be stabilized which is
consistent with the observation that hydrophobic particles
form compact, dense layers at the oil-water interface.11,15
It is also consistent with the conclusions drawn from
ellipsometric measurements of these particles spread at
the planar toluene-water interface which are that the
particles tend to adsorb such that more than a single layer
of particles forms at the interface.16
(ii) Effect of Oil/Water Ratio. In continuous emul-
sification experiments, an emulsion was prepared by the
sequential addition of oil to water (containing a fixed
concentration of particles, 0.7 wt %). The variation in the
average drop diameter and uniformity of the emulsion
drop size distribution as a function of φo are summarized
in Figure 6. As φo increases from 0.1 to 0.6, the average
drop diameter of the water continuous emulsions increases
by a factor of approximately 6, while their uniformity
decreases. Below an oil volume fraction of 0.67, the
emulsion dispersed in water, while at higher oil volume
fractions the emulsion dispersed in oil. Catastrophic phase
inversion therefore occurs at this oil volume fraction. For
the water-in-oil (w/o) emulsions, as the oil volume fraction
was increased, the average drop diameter decreased while
the uniformity of the distribution remained relatively
constant. Thus, the drop size of these solid-stabilized
emulsions tends to a maximum at inversion, unlike for
surfactant-stabilized systems where the drops tend to a
minimum size at inversion.
Figure 7 presents the drop size distributions of the water
continuous emulsions for different oil volume fractions.
As the oil volume fraction increases, the distribution
changes from a single broad log-normal distribution at φo
) 0.1 to a slightly bimodal distribution at φo ) 0.4. The
(15) Yan, N.; Masliyah, J. H. Colloids Surf., A 1995, 96, 229.
Binks and Whitby
Figure 7. Drop size distributions for emulsions of 350 cS PDMS
and 10-2 M NaCl stabilized by 0.7 wt % hydrophobic silica in
the aqueous phase formed as the oil volume fraction was
increased continuously between 0.1 (o/w, dotted line), 0.4 (o/w,
dashed line), and 0.6 (w/o/w, full line). In the multiple emulsion,
the size of the oil globules was measured. Notice the narrowing
of the distribution with increasing oil content.
drop population becomes biased toward the larger drops
as it shifts toward the larger diameters and the remaining
fraction of smaller drops becomes a small separate
population which eventually disappears. At φo ) 0.5 and
0.6, the drop size distribution is a single narrow log-normal
distribution. Since the particle concentration in the system
falls progressively, it seems that the addition of extra oil
increases the interfacial area which may not be sufficiently
covered by the available particles. Thus, the drop size
increases as the drops coalesce. Remarkably, the most
monodisperse emulsions (U ≈ 0.2) are those possessing
the largest drop size.
Monodisperse surfactant-stabilized emulsions have
been prepared in a variety of ways. These include the
size-based separation of polydisperse emulsions using
micelle-induced depletion interactions,17 the rupturing of
emulsions with large drop sizes under controlled condi-
tions,18 and the extrusion of the phase to be dispersed
through membranes,19 capillaries,20 or channels21 into a
flowing continuous phase containing stabilizers. Solid-
stabilized monodisperse emulsions have been prepared
by replacing the surfactant molecules in a surfactant-
stabilized monodisperse emulsion by particles22 and using
the phenomenon of limited coalescence.23,24 In the latter
process, emulsions with high dispersed phase volume
fractions are prepared with insufficient particles available
to stabilize the drops initially formed. The drops subse-
quently coalesce together, reducing the interfacial area,
with smaller drops coalescing more rapidly. Thus, a narrow
distribution of drop sizes is produced. The experiments
(16) Binks, B. P.; Clint, J. H.; Dyab A. K. F.; Fletcher, P. D. I.; Kirkland,
M.; Whitby, C. P. Langmuir 2003, 19, 8888.
(17) Bibette, J.; Roux, D.; Nallet, F. Phys. Rev. Lett. 1990, 65, 2470.
(18) Mabille, C.; Schmitt, V.; Gorria, Ph.; Leal Calderon, F.; Faye,
V.; Deminiere, B.; Bibette, J. Langmuir 2000, 16, 422.
(19) Omi, S. Colloids Surf., A 1996, 109, 97.
(20) Umbanhowar, P. B.; Prasad, V.; Weitz, D. A. Langmuir 2000,
16, 347.
(21) Liu, X.; Nakajima, M.; Nabetani, H.; Xu, Q. Y.; Ichikawa, S.;
Sano, Y. J. Colloid Interface Sci. 2001, 233, 23.
(22) Giermanska-Kahn, J.; Schmitt, V.; Binks, B. P.; Leal Calderon,
F. Langmuir 2002, 18, 2515.
(23) Wiley, R. M. J. Colloid Sci. 1954, 9, 427.
(24) Arditty, S.; Whitby, C. P.; Binks, B. P.; Schmitt, V.; Leal-Calderon,
F. Eur. Phys. J. E. 2003, 11, 273.
Silica Particle-Stabilized Emulsions Langmuir, Vol. 20, No. 4, 2004 1135

described here follow the same method as that developed

by Binks and Rodrigues25 recently to prepare monodis-
perse emulsions of triglyceride oil-in-water, which is
related to limited coalescence. It is suited to producing
concentrated emulsions of relatively large drops with a
narrow size range and, in the emulsions produced here,
it is possible to dilute the emulsion (to an oil volume
fraction of 0.2) by gently stirring it into aqueous 10-2 M
NaCl without altering the drop size distribution.
The drop size distributions shown above are volume
weighted distributions, expressed in terms of equivalent
spheres (measured by light scattering). The drops were
also examined visually by microscopy. In the images shown
in Figure 8, it can be seen that some of the oil drops are
not spherical, which is attributed to the high viscosity of
the oil. At low oil volume fractions, a simple o/w emulsion
forms. However, at φo of 0.5 and 0.6, the oil globules contain
within them smaller drops of (presumably) water, indi-
cating a multiple water-in-oil-in-water (w/o/w) emulsion
has formed. There was no evidence of creaming of the oil
globules in the viscous emulsion produced at these high
oil volume fractions. There was also no evidence of
coalescence of the oil globules to produce a bulk layer of
oil above the emulsion.
The formation of multiple emulsions typically requires
two different surfactants26 or two types of solid particles
(with different hydrophobicities)27 to stabilize the oil-
water interfaces of opposite curvature. Unlike surfactant-
stabilized multiple emulsions which are inherently un-
stable, solid-stabilized multiple emulsions are very stable
to coalescence of both the inner drops and outer globules
because of the strong adsorption of the particles to the
oil-water interface.27 There has only been one reported
case of a multiple emulsion stabilized by a single type of
solid particle,25 and it is unclear how stabilization of both
water-in-oil and oil-in-water drops can be understood on
the basis of the current model used to relate emulsion
type to the particle wettability. In the emulsions prepared
here, however, the high viscosity PDMS oil is a mixture
of silicone oils of different viscosity. Variations in the
wettability of hydrophobic silica surfaces for silicone oils
of different viscosity have been observed.28 This together
with the possibility of the adsorption of PDMS onto silica
surfaces (the extent of which may also vary for polymers
of different molecular weight) may provide some clues as
to the origin of the formation of multiple globules.
The emulsion formed above inversion is a viscous water-
in-oil one. Although the average drop diameter of the
emulsion decreases with increasing oil volume fraction,
the drop size distributions are typically very broad,
consisting of a large peak at high diameters with several
shoulders at smaller diameters. This is consistent with
the large range of drop sizes observed in optical microscopy
images of the emulsions, some of which are shown in
Figure 9. Interestingly, the majority of the water drops
are spherical in shape, indicating that unlike the viscous
oil drops, these drops are able to relax back into spheres
once the distortion caused by homogenization is halted. Figure 8. Optical micrographs of emulsions of 350 cS PDMS
(iii) Effect of Emulsification Time. Under conditions and 10-2 M NaCl stabilized by 0.7 wt % hydrophobic silica in
away from phase inversion, if it occurs, it is typically the aqueous phase (prepared by the continuous addition of oil)
observed that as two immiscible liquid phases are at oil volume fractions of 0.3 (o/w, upper), 0.4 (o/w, middle), and
homogenized together, the average drop diameter de- 0.6 (w/o/w, lower). The bars represent 50 µm.

(25) Binks, B. P.; Rodrigues, J. A. Langmuir 2003, 19, 4905.
(26) Garti, N. Colloids Surf., A 1997, 123-124, 233.
(27) Binks, B. P.; Dyab, A. K. F.; Fletcher, P. D. I. In Proceedings of
3rd World Congress on Emulsions; CME: Boulogne-Billancourt, 2002;
1-N°‚1, p 10.
(28) Binks, B. P.; Lumsdon, S. O. Phys. Chem. Chem. Phys. 2000, 2,
2959.
creases with mixing time, until insufficient energy is
supplied by the agitation to overcome the barrier to further
drop deformation (see, for example, refs 29-32). For a
batch emulsion at a low oil volume fraction (0.33), far
(29) Rehfeld, S. J. J. Colloid Interface Sci. 1967, 24, 358.
1136 Langmuir, Vol. 20, No. 4, 2004
Figure 9. Optical micrographs of a 10-2 M NaCl in 350 cS
PDMS emulsion stabilized by 0.7 wt % hydrophobic silica in
the aqueous phase (prepared by the continuous addition of oil)
at oil volume fractions of 0.7 (upper), 0.8 (middle), and 0.9
(lower). The bars represent 200 µm.
from the conditions required for phase inversion, this
decrease in average drop diameter with increasing emul-
sification time is observed as shown in Figure 10. In Figure
11, the drop size distributions of the emulsions after
particular emulsification times are shown. Initially, the
(30) Groenweg, F.; van Dieren, F.; Agterof, W. G. M. Colloids Surf.,
A 1994, 91, 207.
(31) Abismail, B.; Canselier, J. P.; Wilhelm, A. M.; Delmas, H.;
Gourdon, C. Ultrasonics Sonochemistry 1999, 6, 75.
(32) Baldyga, J.; Bourne, J. R.; Pacek, A. W.; Amanullah, A.; Nienow,
A. W. Chem. Eng. Sci. 2001, 56, 3377.
Binks and Whitby
Figure 10. Effect of homogenization time on the mean drop
diameter of a 33 vol % 350 cS PDMS in 10-2 M NaCl emulsion
stabilized by 0.7 wt % hydrophobic silica in the aqueous phase
(batch prepared).
Figure 11. Drop size distributions for 33 vol % 350 cS PDMS
in 10-2 M NaCl emulsions stabilized by 0.7 wt % hydrophobic
silica in the aqueous phase (batch prepared) formed after
emulsification times of 1 s (full line), 30 s (dashed line), and 180
s (dotted line).
drop size distribution consists of a peak with a shoulder
at smaller diameters. With increasing emulsification time,
the shoulder disappears and the peak shifts to lower
diameters as the drop size approaches the minimum
possible. The stability of the emulsion to creaming does
not change significantly as a function of homogenization
time except after long times where it increases. Presum-
ably under these conditions, which are far from inversion,
the drop size is controlled by both the amount of energy
supplied for fragmentation and limited coalescence.
For a batch emulsion at a higher oil volume fraction
(0.6), however, close to conditions required for phase
inversion, it is observed that the average drop diameter
of the emulsion increases with emulsification time as
shown in Figure 12. Because of the higher ratio of oil to
particles, larger drop sizes would be predicted and the
high volume fraction of oil in the emulsion must increase
the frequency of drop collisions during homogenization
and thus the extent of coalescence. The uniformity of the
drop size distribution decreases rapidly over about the
first 20 s of emulsification before becoming approximately
Silica Particle-Stabilized Emulsions
Figure 12. Effect of homogenization time on the mean diameter
(left-hand ordinate, squares) and uniformity (right-hand or-
dinate, triangles) of 350 cS PDMS-10-2 M NaCl batch emulsions
at φo ) 0.6 stabilized by 0.7 wt % hydrophobic silica in the
aqueous phase. After 2 min of emulsification, the emulsion
catastrophically phase inverts to an oil continuous one (dotted
line).
constant with emulsification time. Again, an inverse
relationship between the average drop diameter and the
uniformity is observed. After about 2 min of homogeniza-
tion, the emulsion catastrophically phase inverts. The
water-in-oil emulsion thus formed sediments rapidly and
almost completely after the mixing is halted. Emulsion
formation close to inversion seems to be dominated by
coalescence, which can lead to inversion if the emulsion
is homogenized for long enough.
This type of behavior has been observed for several
surfactant-free or surfactant-containing systems. It has
been attributed to the inclusion of drops of the continuous
phase into coalescing drops of the dispersed phase, which
increases the effective volume fraction of the dispersed
phase and thus leads to phase inversion if a critical packing
fraction of the globules is reached during emulsification.9,10
The inclusion mechanism obviously leads to the formation
of multiple emulsions, and the change in the structure of
multiple emulsions formed in this way with emulsification
time has also been recently studied.33 At lower oil volume
fractions, it is argued that although inclusion occurs the
rate of escape (or dissolution) of the inner drops is much
higher, so under these conditions inversion is not observed
even for long emulsification times. Certainly, in the system
described here, the formation of multiple emulsion drops
was only observed at high oil volume fractions. Although
it seems clear that at high oil volume fractions the inclusion
mechanism occurs because of coalescence of drops (whether
two or many), it has recently been proposed33 that where
inclusion is observed at low dispersed phase volume
fractions it may occur during the deformation of drops.
(33) Sajjadi, S.; Zerfa, M.; Brooks, B. W. Chem. Eng. Sci. 2002, 57,
663.
Langmuir, Vol. 20, No. 4, 2004 1137
Part of the drop may become concave toward the drop
phase trapping the continuous phase in the concave area
as the surface expands.
The effect of emulsification time on emulsions formed
by the continuous variation of oil volume fraction was
also investigated. An emulsion with φo ) 0.3 was prepared
by continuous addition of oil (with re-homogenization steps
of 1 min) and then enough oil was added to increase the
oil volume fraction to 0.4. This mixture was homogenized
together and the average drop diameter was monitored
as a function of emulsification time. The average drop
diameter did not change significantly even after 20 min
of homogenization. Clearly, a balance between the frag-
mentation and coalescence processes is rapidly achieved.
Yet, this is not entirely the case. Significantly, for
emulsions stabilized by higher particle concentrations (2
wt %) that are prepared by continuous addition of oil, it
is necessary to re-homogenize for shorter times (30 s) after
each addition of oil to form emulsions with narrow drop
size distributions at high oil volume fractions. This also
increases the oil volume fraction at which phase inversion
occurs. At high oil volume fractions, very large volumes
of viscous oil are being added to increase the oil volume
fraction by the required amount. Obviously, depending
on the particle concentration and presumably the volume
of oil being added, there is an optimum re-homogenization
time, during which just enough energy is supplied to
fragment the bulk oil. Presumably, the drops formed by
fragmentation are not well stabilized by particles and thus
will coalesce with emulsion drops previously formed. For
longer homogenization times, there is greater opportunity
for fragmentation of some of the drops to occur and a
monodisperse emulsion is not prepared close to inversion.
Conclusions
We have investigated the influence of particle concen-
tration, oil volume fraction, and emulsification time on
the formation, structure, and stability of emulsions of
silicone oil and water stabilized by partially hydrophobic,
monodisperse silica particles. As the particle concentration
in the emulsions is increased, the average drop diameter
decreases until a minimum size is reached as the extent
of coalescence during drop formation is reduced. Increasing
the oil volume fraction at fixed particle concentration
causes an increase in the average drop diameter until
catastrophic phase inversion occurs. Phase inversion is
associated with the formation of multiple w/o/w emulsions,
of very low polydispersity, as coalescence dominates
emulsion formation. At lower oil volume fractions, the
drop size of simple emulsions decreases with increasing
emulsification time as emulsion formation involves both
fragmentation and limited coalescence.
Acknowledgment. This work was funded by Rhodia
Services (France). The authors thank M. Kirkland (Uni-
lever Research Colworth) for performing the LTFESEM
measurements.
LA0303557