Upload
emily-amancio
View
216
Download
0
Embed Size (px)
Citation preview
8/3/2019 Surf Act Ant Free Emulsion
1/13
Colloids and Surfaces
A: Physicochemical and Engineering Aspects 180 (2001) 4153
Surfactant-free O/W emulsion formation of oleic acid andits esters with ultrasonic dispersion
Keiji Kamogawa a,b, Hidetaka Akatsuka c, Mitsufumi Matsumoto c,Shoko Yokoyama d, Toshio Sakai c, Hideki Sakai b,c, Masahiko Abe b,c,*
a Ele. & Sec. Ed. Bureau, The Ministry of Education Science, Sports and Culture, Science, Kasumigaseki, Chiyoda,
Tokyo 100-0013, Japanb Institute of Colloid and Interface Science, Science Uni6ersity of Tokyo, 1-3 Kagurazaka, Shinjuku, Tokyo 162-8601, Japan
c Faculty of Science and Technology, Science Uni6ersity of Tokyo, 2641 Yamazaki, Noda, Chiba 278-8510, Japand
Kyoritsu College of Pharmacy, 1-5-30 Shibakoen, Minato, Tokyo 105-8512, Japan
Received 14 January 2000; accepted 5 September 2000
Abstract
Dispersibility and stabilizing factors for surfactant-free O/W emulsion were investigated with oleic acid (OA) an
its esters, focusing on the effects of their weak polarity, molecular length and branched chain structure, in compariso
to normal hydrocarbons. The droplet size distributions obtained by the dynamic light scattering method appeared t
be discrete but almost singly peaked except for OA. For OA monoesters, the droplet growth was found to b
continuous and retarded as the ester chain length increased, in contrast to the discrete, fast growth in OA dispersion
In the case of glycerol trioleate (GTO), a branched ester, aqueous dispersions of extremely fine droplets could bprepared and the number distribution of droplet diameters showed a single peak in the nanometer range. This hig
dispersibility remained unchanged for about a year after preparation to give the dispersions a good stability. Chang
of the observed x potential, Fourier transform-infra red (FT-IR) spectrum, fluorescence spectrum of probes indicate
that a particular carboxyl acid group network is formed in the droplet sphere to make it more stable than expecte
while the interior of oil droplets is hydrophobic. The x potential change, in particular, was found to be high
correlated with these of the carboxyl CO stretching frequency and the reciprocal droplet diameter. The dropl
stability evaluated by increase in the diameter revealed a biphasic growth consisting of fast and slow modes. The fa
growth at early stages (in hours) observed for OA and methyl oleate (MO) was found to proceed by the Ostwa
ripening mechanism through rate analysis. On the other hand, the slow growth at later stages (in days) found for th
other esters showed a semi-logarithmic dependence on the oil viscosity. This seems to be caused by an Arrhenius-typ
activation factor in the stepwise flocculation/coalescence rate. 2001 Elsevier Science B.V. All rights reserved.
Keywords: Surfactant-free emulsion; Oleic acid and its esters; Growth mechanism
www.elsevier.nl/locate/colsur
* Corresponding author. Tel.: +81-471-248650; fax: +81-471-248650.
E-mail address: [email protected] (M. Abe).
0927-7757/01/$ - see front matter 2001 Elsevier Science B.V. All rights reserved.
P I I : S 0 9 2 7 - 7 7 5 7 ( 0 0 ) 0 0 7 5 8 - 5
8/3/2019 Surf Act Ant Free Emulsion
2/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 415342
1. Introduction
When an oil-in-water type (O/W) emulsion isprepared using two hardly miscible components,e.g. water and a hydrocarbon, a suitable surfac-tant is usually used to enhance emulsification andmake the emulsion formed stable. O/W type
emulsions are widely employed in various prod-ucts such as medicines like highly nutritiousagents and lipid microspheres [1,2], cosmetics [3],and foods [4]. While ultrafine particles of inor-ganic and organic substances are usually preparedin the presence of surfactant [5], the removal ofadsorbed surfactant molecules from the surface ofultrafine particles is difficult. Correspondingly, in-troduction of surfactant-free emulsions is desiredsince the droplets are composed of oil moleculesalone. Transient dispersions are familiar to us, aswell known by oil-and-vinegar sauce, solvent ex-
traction, etc. Droplets in such emulsion are ex-pected to have a clear surface and the interiorwith the highly hydrophobic nature of oil proper.
In view of these, we have prepared droplets inwater of aromatic hydrocarbons like benzene [6],straight chain hydrocarbons like n-hexane andn-hexadecane [7], and branched chain hydrocar-bons like squalane by ultrasonic irradiation andexamined the dispersibility and the stability of oildroplets formed. We then found that dispersionsof such oil droplets are regularly characterized bydiscrete size distributions with a certain degree of
stability [8]. The behavior may be classified into(1) benzene type giving small-sized, unstabledroplets, (2) n-hexane type giving middle-sized,stable droplets and (3) n-hexadecane type givingsmall-sized, stable droplets. In addition, mixing ofn-hexadecane with benzene was found to suppressdroplet growth remarkably, despite the fact thatthe interfacial tension (a) increased on mixing.This would mean that a decrease in a is notnecessarily coupled with the dispersion stability ofsurfactant-free emulsions. Among the other prop-
erties, the branched structure possessed bysqualane and surface charging supposed for n-hexadecane seem to be promising factors [9]. It isthus worth studying the dispersibility and thestability of such oils that have a branched struc-ture as well as a polar group instead of inducedcharges.
As to the kinds of oils to be used, fatty acid
and their glycerol esters have weakly pol
groups, in addition to their long hydrophob
chains and branched structure. They are too muc
hydrophobic to be regarded as surfactant as rep
resented by a very low HLB value of 1 an
solubility in water. Such a low HLB value mean
that they may act only as weak emulsifiers suiable for W/O dispersion. In a dynamic sens
however, we can expect a rather high stability
water of their droplets. In addition, oleic acid
present as a component in complex lipids, main
phospholipids of biomembranes, and thus it pla
an important role in various tissues and organ
while diversely changing its chemical and ster
structure and molecularly assembling mode [10
In the present study, we examine the dispersibilit
and growth characteristics of methyl oleate, dec
oleate, oleyl oleate, glycerol trioleate, and ole
acid as the control to find the conditions nece
sary for preparing surfactant-free emulsion of
long term dispersion stability.
2. Experimental
2.1. Materials
Methyl oleate (MO, NOF Corp., 99.7% purity
decyl oleate (DO, NOF Corp., 98.5% purity
oleyl oleate (OO, NOF Corp., 91.6% purity), glyerol trioleate (GTO, NOF Corp., 97.0% purity
oleic acid (OA, NOF Corp., 99.7% purity) we
used as oils as supplied. The residual impuriti
were estimated to be unsaturated fatty acids wit
different chain lengths or their esters. Diole
esters might be present at about 3.0% of th
glycerol esters, which is sufficiently low not t
activate the interface. Distilled and deionized w
ter for injection (Ohtsuka Pharmaceutical Co
was employed without further purification. Pyren
(Wako) and 8-anilino-1-naphthalene sulfonic ac(ANS, Wako) were used as fluorescence probes
2.2. Preparation of oil droplet dispersions
Preparation of oil droplet dispersions was pe
formed in the following way. A mixed liquid o
8/3/2019 Surf Act Ant Free Emulsion
3/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 4153
water and oil at a given mixing ratio (50 ml)
was pretreated by stirring for 1 min in a
vortex mixer and it was then ultrasonically irra-
diated for 8 min on an ultrasonic cleaner (Bran-
sonic 220, 125W, Smith-Kline Co.). The
exposure time of 58 min was necessary for the
size distribution in the nmmm region to appear
steadily.
2.3. Measurement of oil droplet size distribution
Measurement of oil droplet size distribution
was performed periodically by a dynamic light
scattering method (homodyne method) using an
NICOMP 380 ZLS (Particle Sizing System Co.).
The oil concentration was 0.5 mM, the light
source was a diode pump solid state laser
(DPSS laser) with a wavelength of 535 nm, and
the measuring angle was 90.
2.4. Interfacial tension and x potential
measurement
The surface properties of oil droplets were
evaluated with interfacial tension and x poten-
tial. The interfacial tension was obtained with a
platinum Wilhelmy plate in Model CBVP-Z in-
terfacial tension meter (Kyowa Interface Science
Co. Ltd.), where samples were kept at 30C be-
fore the measurements. Data were stored whenthe value became invariant for more than 10
min.
x Potential was measured on oil droplets in
dispersions prepared by the method described in
2.2 with the laser Doppler method (heterodyne
method). This method gives the electrophoretic
mobility and x potential of fine particles moving
in an electric field based on the Doppler shift, a
frequency shift of laser light scattered by the
particles. The oil concentration and the light
source were the same as those in the dropletsize distribution measurements. The laser output
power was 10 mW, the scattering angle was
19.8, and the electric field strength was 5.0 V
cm1. Sensitivity test was done for the system
of n-hexadecane droplets in water, which was
found to be highly stable. It gave a value of
70 mV, in agreement with the literature da
for docosane particles [11].
2.5. FT-IR spectrum measurement
Fourier transform-infra red (FT-IR) spectru
measurement was conducted on oil droplet di
persion using the ATR method at an oil concentration of 10 mM to further evaluate the surfac
properties of the droplets. The apparatus use
was a spectrophotometer JIR-5300 (JEOP Co
equipped with an ATR cell (Model TNL-13
Tunnel Cell) (JEOP Co.). The cell is made
zinc selenide (ZnSe), an IR-transparent materi
(crystal).
2.6. Fluorescence spectrum measurement
Fluorescence spectrum measurement wmade using two fluorescence probes, pyrene an
ANS, to investigate the environment of o
droplets. The apparatus employed was
fluorospectrophotometer RF-5000. (Shimadzu).
2.6.1. Measurement using pyrene
Fluorescence spectrum measurement usin
pyrene as the probe was carried out on a dispe
sion of oil droplets containing 1 mM pyren
prepared according to the procedures de
cribed in Section 2.2. The oil concentration the dispersion was varied from 0.1 to 3
mM. The ratio, I1/I3, for pyrene monomer w
calculated from the fluorescence spectrum o
tained with freshly prepared dispersion (Ex 33
nm).
2.6.2. Measurement using ANS
In spectral measurements using ANS as th
probe, the fluorescence spectrum of ANS di
solved in oil at 0.1 mM was compared with tha
in droplets of oil containing 0.1 mM ANS dipersed at 5 mM in water (Ex 337 nm)
Preparation of oil droplet dispersion and me
surements of droplet size distribution, x poten
tial, and fluorescence spectrum were performe
at 30C. FT-IR spectra were measured at 2292C.
8/3/2019 Surf Act Ant Free Emulsion
4/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 415344
3. Results and discussion
3.1. Initial dispersibility and growth
Fig. 1 shows a typical size distribution of oleic
acid (OA) droplets dispersed in water.
OA droplets were found to have a discrete size
distribution immediately after preparation of dis-persion with two size classes, S and M, the former
Fig. 2. (A) Stepwise coalescence model and (B) critical flocc
lation/coalescence model.
Fig. 1. Particle size distribution of OA droplets in water in
number% (solid curve) and in volume% (broken curve).
Droplets were prepared at 0.5 mM concentration. The arrows
S, M and L denote small (S), middle (M) and large (L) classes,
respectively.
appearing at 70100 nm, and the latter at 300
700 nm. These peaks were distinguishable in e
ther number% or volume%, with a highsensitivity for small or large droplets, respectivel
As time elapsed, the droplets grew in such a wa
that the height of peak changes while the pea
positions for S and M classes remain almo
unchanged and L class appeared at 1000 250
nm 3 days later. Extremely large droplets abov
30005000 nm observed 10 and 60 min aft
presentation, seem to be partially coalesced drop
which disappear through creaming and will not b
discussed further. We also noticed that the S-cla
size of OA droplets is much smaller than 200 nmthe size found for oil droplets by mechanic
dispersion with 0.1 mol l1 of SDS [12]. Th
growth of OA droplets is much faster than that o
n-hexadecane (C16H30) droplets [7], which have
comparable size to OA with respect to carbo
number. As a whole, the behavior of OA drople
much resembled that found for benzene drople
[6], over that for n-hexadecane droplets, even th
acid has a carboxylic acid group and a long alk
chain. This resemblance suggests the presence of
critical flocculation/coalescence scheme in O
droplet growth, as previously assumed for be
zene droplets [6,8]. Basically, small oil droplets a
expected to grow through several processes, in
cluding coalescence and the Ostwald ripening.
As shown in Fig. 2(A), the former is unde
stood to proceed by stepwise collision proces
8/3/2019 Surf Act Ant Free Emulsion
5/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 4153
resulting in gradual increase in the droplet size as
has been found elsewhere. When droplets are
sufficiently resistive to collision or collision-in-
duced coalescence, however, a certain degree of
flocculation would be required to open an effec-
tive fusion path to coalescence as shown in Fig.
2(B). This critical flocculation/coalescence scheme
can be related with discrete distribution andgrowth of droplet size, which are characteristic for
the surfactant-free hydrophobic oil droplets [8].
Methyl oleate droplets (MO), by contrast, had
almost a single peak of S class in their size
distribution (Fig. 3(a)) and grew more rapidly.
Thus, the droplets showed a growth process in
which their size distribution changed continu-
ously, keeping the bimodal distribution in S an
M classes. Both of S class and M class droplets o
decyl oleate (DO) grew in size from about 90 an
400 nm to about 130 and 700 nm, respectivel
Such continuous growth is rarely observed fo
surfactant-free emulsions though it is quite popu
lar in the surfactant-assisted emulsions [13]. Suc
continuous size drift suggests a significant contrbution to droplet growth of stepwise flocculatio
and coalescence referred in Fig. 2. Correspon
ingly, this hybrid profile is somewhat differe
from the characteristics of normal hydrocarbo
droplets.
In Fig. 3(b and c) are shown the droplet siz
distributions for DO (b) and OO (c), both havin
a longer alcoholic alkyl chain than methyl oleat
Freshly prepared droplets of these two esters als
showed a single peak in S class in their siz
distributions as in the case of MO droplets. The
growth also followed a similar path, continuous
changing their size distribution from the one in
class to the other in M class. Droplets in L siz
class were negligibly few in the number% distribu
tion. While it took only 23 days for M
droplets to grow two to three times in size, 2
weeks were needed for DO and OO droplets t
become two to three times larger in size. In othe
words, the rate of droplet growth slows down a
the ester chain length increases.
Fig. 3(d) shows the size distribution for GT
droplets, which demonstrates that the S class sizdroplets observed immediately after dispersio
preparation remain unchanged for at least a yea
This is the longest period during which oil dropl
dispersion keeps its initial stability in our surfa
tant-free emulsion study. Such a long term stabi
ity would be due partly to the fact that GTO is
branched ester with long hydrophobic group
whose branched structure extends to the who
molecule and partly to a high viscosity of th
ester as mentioned later.
The peak diameter for S class droplets wplotted against time for esters in Fig. 4. Since th
growth continued monotonously for a week, th
size of MO droplets reached 500 nm, which corr
sponds to M class size. From the curves, th
growth speed was roughly estimated for S cla
droplets as 60, 10, 4 , and 0.5 nm per day for MO
Fig. 3. Particle size distribution of oil droplets in water in
number % for (a) MO, (b) DO, (c) OO, and (d) GTO.
Droplets were prepared at 0.5 mM concentration. The arrows
S, M and L denote small (S), middle (M) and large classes (L),
respectively.
8/3/2019 Surf Act Ant Free Emulsion
6/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 415346
Fig. 4. Changes in droplet diameter as a function of time. ()
MO, () DO, (2) OO and () GTO. Data were taken fromFig. 3 as the average in number % distribution.
mPas) it failed to give sufficiently small drople
in water, suggesting no direct correlation betwee
droplet dispersibility and viscosity for the aci
This would mean that the initial dispersibility
determined by the specific nature of carboxyl
acid group at the interface or mechanically r
stricted by the high viscosity and a given cavita
tion field in the ultrasonic bath.
3.2. Surface properties
In view of this, we attempted to see if an
relation exists between the dynamic stability of o
droplets and their surface properties in the growt
process. Fig. 6 shows the x potential of o
droplets dispersed in water as a function of tim
for the five oils, indicating that oil droplets have
x potential of 33 to 30 mV soon after dispe
sion preparation.Note here that oil droplets are negative
charged though no electrically charged third sub
stance is present in the dispersion. The initi
levels were deep enough to allow better dispersio
stability for the droplets. As to why the x poten
tial of oil droplets is negative, the following even
that are possible to happen on the droplet surfac
would be responsible.
DO, OO, and GTO esters, respectively. No sig-
nificant difference in interfacial tension (a) was
found among the oils used in the present work.
The avalues for OA/water and GTO/water were
in agreement with the literature values of 15.6 and
23.2 mNm1, respectively [14]. The values were
high enough (1330 mNm1) to prevent sponta-
neous emulsification. On the other hand, the vis-
cosity increased with increasing alkyl chain lengthof alcohol group in an increasing order of MO
(4.88 mPas) BDO (12.7 mPas) BOO (24.0
mPas) BGTO (56.0 mPas). Hence, there is a
close relationship between droplet dispersibility,
i.e. the smallness in droplet size, and viscosity for
the four esters studied. Droplet size is thought to
be determined by interfacial tension (a) and vis-
cosity (p) of the fluid under a shear stress [15,16],
occasionally by the ratio of a/p [17]. This rela-
tionship was confirmed recently for surfactant-
free droplets of n-alkanes in water [8]. The initial
droplet size in Figs. 1 and 3 was then plotted
against a/p in Fig. 5, as is proposed for surfac-
tant-added emulsions.
The initial size was independent of or slightly
decreasing with a/p for the oils under the present
condition. Although OA has a high viscosity (23.0
Fig. 5. Dependence of initial peak size of droplets on a/p
pure oil. () OA, () MO, () DO, (2) OO and () GT
The sizes were taken from Figs. 1 and 3.
8/3/2019 Surf Act Ant Free Emulsion
7/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 4153
Fig. 6. Changes in x potential of oil droplets dispersed inwater as a function of time. () OA, () MO, () DO, (2)
OO, () GTO.
appreciable growth of OA and MO. On the othe
hand, the drift of x potential value might b
related with the diameter increase. Fig. 7(a) show
the plot of x potential change against diamet
has no linear correlation but hyperbolic one. Th
plot was then made against the reciprocal diam
ter to yield a linear relation as shown in Fig. 7(b
This linear relationship does not seem to arisfrom conservation of the total charges on th
decreasing total droplet surface area. Since th
droplet growth at a constant total volume shou
decrease the total surface area by 1/diameter, th
charge will be concentrated on the droplet surfac
to raise the negative x potential in proportion t
diameter. A plausible source for changing x p
tential is variable charges induced on the surfac
by the intrinsic droplet properties. The linear rel
tionship may then suggest a geometrical expan
sion or cracking at the droplet surface, whic
would magnify adsorption probability of the r
sponsible ions onto the surface. This change in
potential with time would then be better related t
the structural changes of oil droplets.
3.3. FT-IR spectra
FT-IR spectrum measurements were conducte
on aqueous dispersions of OA, MO and GTO i
order to examine the surface of oil droplets
detail with emphasis on the state of carboxyl
acid groups (Fig. 8).The spectra are shown in the region from 165
to 1850 cm1 since peak shifts were observed i
the absorption range corresponding to the C
stretching vibration. When the spectra for oi
alone (i) were compared with those temporal one
for their droplets in water (iiiv), the peak at 174
cm1 for glycerol trioleate (a) was found to shi
by 5 cm1 to the high wave number side for i
droplets immediately after dispersed in water (i
and no further change was observed in the spe
trum even after a week (iv), while an upward shiof absorption peak was also observed for ole
acid (b) and methyl oleate (c) from 1708 (i) t
1712 cm1 (ii) and from 1740 (i) to 1744 cm
(ii), respectively, but the shifts gradually dimin
ished as time elapsed, making the absorptio
peaks return to the initial positions. Since th
1. Carboxylic acid groups that have a permanent
dipole and are weakly polar (C+O) orient
on the surface of oil droplets to induce a
negative charge around them.
2. Hydroxyl ions generated in the dissociation of
water are stabilized on the droplet surface due
to the difference in dielectric constant between
the oil droplet and aqueous phase [18], and
3. The permanent and induced dipoles are or-
dered following the structuring of hydrophobi-
cally hydrated water molecules.
Comparison of x potential changes with time
among the oils revealed that GTO droplets with a
good stability keep a highly negative x potential
of33 mV over a week while the x potentials of
OA and MO droplets with a poor stability rise
from a similar value with time up to 10 and
18 mV for the former and the latter, respec-
tively. These initial x potential values in Fig. 6 are
so similar to each other that they can not be
correlated with the difference in the growth speed
evaluated from the initial slope in Fig. 4, where
OA exhibits a pronounced slope of Dd/Dt, much
larger than that for MO. The initial x potential
value is then no useful for predicting the fast or
8/3/2019 Surf Act Ant Free Emulsion
8/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 415348
absorption due to the CO stretching vibration of
oleic acid generally appears at 1708 cm1 for its
dimers and it shifts to a higher wave number of
1760 cm1 for its monomers [19], the observed
upward shift of the absorption peak would be
caused by breakdown of CO/CO interaction
between neighboring molecules when the oil is
dispersed as droplets in water. Hence, a particularorder, which differs from the structure of
oil itself, is presumably formed in droplets for the
oleic acid esters. Invariance of the upward shift
suggests that formation of the particular order
should be a stabilizing factor and the order is
protected by the presence of alcoholic groups. The
return of the peak to the initial position observed
for OA and MO would be brought about by
diminution of the particular structure throug
molecular diffusion, droplet coalescence, wat
penetration and carboxylic acid group migratio
to the surface, etc. toward a reestablishment
the molecular arrangement in the bulk of o
Then, an attempt was made to correlate th
CO shifts directly to the x potential changes Fig. 9.
A correlating curve was obtained for each
OA and MO. Thus, the two curves similarly ha
a negative slope and the upper one tended t
converge to 17 mV. The similarity suggests th
the drift ofx potential is mainly controlled by th
state of the carboxylic acid groups.
Fig. 7. Correlations of x potential (a) with droplet diameter () fitted to a hyperbola and (b) with reciproca1 diameter () fitt
to a line at corresponding time. Data were taken from Figs. 4 and 6.
8/3/2019 Surf Act Ant Free Emulsion
9/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 4153
Fig. 8. IR spectra in the carboxyl CO stretching region for (a) GTO, (b) OA and (c) MO. (i) Oil phase, (ii) oil-droplets just aft
preparation, (iii) after 3 days and (iv) after 7 days.
3.4. Fluorescence probe
The particular structuring related with the car-
boxylic acid group may also be responsible for the
hydrophobicity or fluidity of the droplet. The
hydrophobic environment of oil droplets was then
examined using pyrene, a fluorescence probe (Fig.
10).
The ratio of the intensity of band 1 to that of
band 3, I1/I3, both bands being characteristic of
the vibrational structure of fluorescence for
monomeric pyrene molecules, is an index for the
hydrophobicity of the microenvironment sur-
rounding pyrene molecules [20,21]. Pyrene fluores-
cence intensity measurements on droplets of OA,
MO, and GTO yielded a constant value (0.75
0.9) of the ratio for the three oils above an
extremely low oil concentration of 0.3 mmol l1.
The values for dispersions of cyclohexane and
benzene, both are nonpolar oils, are 0.8 and 1.0,
respectively [22], and 1.0 1.2 for SDS micelles
[21,2325]. The values obtained for the oils in this
work locate between those for SDS and dodecane
(0.6) environment [25]. A slight degree of hydra-
tion might be suggested for the dispersed oil
droplets in view of the I1/I3 values of 0.49, 0.53
and 0.52 for bulk OA, MO and GTO, respec-
tively. The values for droplets obtained above are
indicative of a rather hydrophobic environment o
these droplets, much more hydrophobic than th
of micellar systems. An interesting fact is that O
has the lowest value, implying a higher hydropho
bicity of OA droplet. This might result in the fa
break down of the particular structure.
Fig. 9. Correlation between the carboxyl CO frequency and
potential at various times. Data were taken from Figs. 6 an
8 for OA (), MO () and GTO ().
8/3/2019 Surf Act Ant Free Emulsion
10/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 415350
Fig. 10. Temporal change of I1/I3 value in the monomer
fluorescence of pyrene dissolved in oil droplets at 1.0 mM
concentration. OA (), MO () and GTO (). Dotted lines
indicate the values for pyrene in water and those in SDSmicelles.
higher value. Hence, we can use the wavelength
the fluorescence intensity maximum and the quan
tum yield of ANS to estimate the polarity of th
environment in which the dye is present. Mor
over, it is known that even though the dye
dissolved in a polar solvent its fluorescence spe
trum makes a blue shift and its quantum yie
rises if the viscosity of the solvent is high [26ANS fluorescence measurements revealed that th
intensity peak of the dye makes a blue shift
about 60 nm when the dye is in OA, MO, an
GTO from that in water. Similar measuremen
were done for droplets of the three oils containin
0.1 mM ANS. A small blue shift of 1 nm wa
found for OA and MO dispersions and a shift o
7 nm and a slight increase in intensity we
observed for GTO dispersion. The noticeable shi
and the larger peak height for GTO dispersio
indicate a higher adsorpability of the probe to th
droplet surface.
3.5. Growth rate analysis
Based on the experimental findings mentione
so far, the growth of oil droplets is quite likely t
be explainable by the diminution of the particula
structure coupled with the state of carboxylic ac
groups. An examination was then conducted t
know how the particular structure diminishes, o
the basis of the mechanisms mentioned prevously. In general, two growth mechanisms a
possible when an emulsion breaks and separat
While fluorescence probe technique is quite sen-
sitive, the sensitivity is site-selective, and hence, it
is regulated by the dissolution characteristic of the
probe. We, therefore, used ANS as another probe,
focusing on its behavior on the droplet surface.
Fig. 11 shows the fluorescence spectra of ANS in
water, oil and dispersion. When the polarity of
solvent decreases the wavelength at the maximum
fluorescence intensity of ANS generally makes a
blue shift by about 60 nm at its maximum and the
quantum yield rises up to an about 200 times
Fig. 11. Fluorescence emission spectra of 0.1 mM ANS at 30C. (a) ANS in oil and water phase, (b) ANS in oil droplets. Wate
(), OA (------), MO ( ) and GTO ( ).
8/3/2019 Surf Act Ant Free Emulsion
11/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 4153
into two liquids physicochemically (1) flocculation
and coalescence of droplets and (2) molecular
diffusion. The latter is called the Ostwald ripening
[27,28] and caused by the increasing solubility of a
substance with decreasing droplet size. Thus,
molecules of the substance of smaller particles
dissolve in the dispersion medium and are ab-
sorbed into large particles through diffusion,thereby causing a change in particle size distribu-
tion. It can be confirmed with drifts in size in the
opposite directions for smaller and larger droplets
as shown in Fig. 1. For OA dispersion, while M
class droplets grew in size from 300 to 400 nm
and L class droplets at 3.5 mm disappeared, S
class droplets shrank from 100 to 50 nm within 1
h. On the other hand, for MO dispersion, S and
M class droplets grew from 100 to 150 nm and
from 300 to 400 nm, respectively, within 3 h. The
Ostwald ripening is then highly probable in the
growth of OA droplets. This scheme was applied
as below for the fast mode of growth while coales-
cence model was tested for the slow mode of
growth.
3.6. Fast growth mode
Lifshitz et al. [29,30] put forward a theory
(LSW theory) for the Ostwald ripening of fine
particles as expressed by the following equation:
=dr3
dt =8DC
kM
9z2RT (1)
where D is the diffusion coefficient of the dis-
persed phase in water, C
the solubility of the
dispersed phase in water, g the interfacial tension
between water and the dispersed phase, M and z
are the molecular weight and density of the dis-
persed phase, respectively, and r is the mean size
of all droplets in the dispersed phase. Kabalnov
[31] and Taylor [32,33] have succeeded in plausi-
bly explaining the Ostwald ripening of liquid
droplets based on this equation. If the dropletgrowth in the present study obeys the equation,
the cube of the mean droplet size should increase
linearly with time. In order to check this point,
the cube of mean radius in volume% representa-
tion was plotted against time for poorly stable
OA and MO droplets and the most stable GTO
Fig. 12. Plot of the cube of droplet radius as a function
time. The averaged radius were obtained as those provided
the volume% representation. () OA, () MO and () GT
droplets (Fig. 12). The plots were made in thperiod of 3 h within which droplets of OA anMO were found to grow to give those in M sizclass. The value of r3 linearly increased rathsteeply with time for OA droplets while the increase in r3 with time for MO droplets wslightly steep and the value of for the ester wamerely one fortieth of that for OA. Almost nchange with time in r3 was found for GT
droplets. This represents a difference in the rate othe Ostwald ripening between OA and the esterThe difference can be partially ascribable to thlower solubility (C
) of the latter. C
was est
mated from the total organic carbon (TOC) concentration and turbidity changes (Fig. 13).
TOC concentration (a) estimated for saturateoleic acid solution was 3 ppm, i.e. 1105 ml1 when the amount introduced to water w3105 mol l1. In addition, the turbidity of thdispersion (b) started to increase at the introduce
amount of 2105
mol l1
, supporting thsolubility of 1105 mol l1. While sucestimation was limited for GTO, the turbidistarted to rise at a lower concentration than Oto suggest a smaller value of the solubility. Thripening rate according to Eq. (1) was almonegligible for GTO as seen in Fig. 12.
8/3/2019 Surf Act Ant Free Emulsion
12/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 415352
Fig. 13. Solubility analysis of OA in water and GTO in water. (a) TOC plot vs. the provided concentration for OA () and (
turbidity vs. the provided concentration for OA () and GTO (). The arrows from the bottom abscissa indicate an interpolate
position as the solubility point of OA.
3.6.1. Slow growth mode
As mentioned previously, the slow growth rate
can be characterized with Dd/Dt as in Fig. 3. In
view of the time needed for the slow rate to be
observed, this slow mode seems to be related with
the flocculation/coalescence mechanism. The slope
ofDd/Dt obtained from Fig. 4 was plotted against
viscosity in Fig. 14 in a semi-logarithmic manner.
A good correlation was seen between them, only
with slight deviation for MO. This type of correla-
tion seems to suggest an Arrhenius-type depen-
dence of flocculation/coalescence rate on
viscosity, such as the case of deformation energy
necessary for droplets at the transition state to
coalesce. The slight deviation in Fig. 14 should
have resulted from the Ostwald ripening analyzed
in Fig. 12, which also be operative even in M
class. The extrapolation for MO indicates that 13
nm per day or 21% of the apparent growth speed
would be due to flocculation/coalescence and the
remaining Ostwald ripening.
The growth process for droplets of OA and the
esters are regulated by the fundamental rate fac-
tors including their molecular solubility and vis-
cosity, etc. Electrostatic repulsion due to x
potential would effectively prevent flocculation/
coalescence as long as the interaction occurs for
similarly sized droplets. However, the repulsio
seems not preventive enough against molecul
processes such as molecular deposition by th
Ostwald ripening and water penetration relatin
to the higher hydrophobicity of OA droplet. I
addition, even though the repulsion should pr
Fig. 14. Dependence of Dd/Dt for the slow growth mode o
the viscosity of the oil. Data were taken from the slope in Fi
4. () MO, () DO, (2) OO, () GTO.
8/3/2019 Surf Act Ant Free Emulsion
13/13
K. Kamogawa et al. /Colloids and Surfaces A: Physicochem. Eng. Aspects 180 (2001) 4153
vent collision-controlled coalescence, it is not oper-
ative enough in much slower flocculation/coales-
cence determined by the Arrhenius-type activation
factor, exp(DE*/RT). Nevertheless, x potential
seems to be a significant indicator for droplet
growth through its 1/d dependence, in which
droplets with varying x potentials are less stable.
4. Conclusions
From the results mentioned so far, OA and the
four esters were found to form surfactant-free
emulsion with different characteristics. OA can be
regarded as an oil giving benzene-like droplets that
are discretely sized and form unstable dispersion in
spite of its C18 chain length. On the contrary, the
dispersion stability of the four oleic acid esters
becomes higher in accordance with their alcoholicalkyl chain length and branched structure. In
particular, GTO droplets of good dispersibility and
stability could be prepared in the absence of
surfactant. This stability is ascribable to formation
of a particular structure around the carboxylic acid
groups. Although x potential value is not predictive
of droplet stability, it is a good indicator for droplet
growth providing a high correlation with reciprocal
diameter. We propose here that particular structure
formation as well as high oil viscosity are the
promising factors for oil droplets to remain stable
in such surfactant-free emulsion against floccula-tion/coalescence.
Acknowledgements
The authors are grateful to the Cosmetology
Research Foundation for its financial support in
this work.
References
[1] H. Sato, S. Ogoshi, S. Usui, Graphic Presentation of High
Calorie Transfusing Liquids, second ed., Igaku-shoin,
1982, p. 44.
[2] M.Y. Levy, W. Schutze, C. Fuhrer, S. Benita, J. Microen-
capsulations 11 (1994) 79.
[3] A.L.L. Hunting, Cosmetics Toiletries 101 (1986) 49 6
[4] M. Endo, H. Sagitani, J. Jpn. Oil Chem. Soc. 40 (1991) 13
[5] M. Abe, H. Nishino, K. Ogino, Sekiyu Gakkaishi 32 (198
151.
[6] K. Kamogawa, T. Sakai, N. Momozawa, M. Shimaza
M. Enomura, H. Sakai, M. Abe, J. Jpn. Oil Chem. Soc. 4
(1998) 159.
[7] K. Kamogawa, M. Matsumoto, T. Kobayashi, T. Sak
H. Sakai, M. Abe, Langmuir 15 (1999) 1913.
[8] K. Kamogawa, M. Abe, Surfactant-free emulsions, iEncyclopedia of Emulsion Technology, Marcel Dekker,
press.
[9] K. Kamogawa, M. Matsumoto, T. Sakai, H. Sakai, M
Abe, in preparation.
[10] M. Suzuki, Fragrance J. 23 (1995) 105.
[11] D.E. Dunstan, D.A. Saville, J. Chem. Soc. Faraday Tran
88 (1992) 2031.
[12] T. Yamamoto, H. Sakai, Y. Sakai, F. Harusawa, K
Kamogawa, M. Abe, Mater. Technol. 18 (2000) 53.
[13] C.M. Miller, J. Venkatesan, C.A. Silbi, E.D. Sudol, M
EL-Aasser, J. Colloid Interf. Sci. 162 (1994) 11.
[14] Y. Minegishi, T. Takeuchi, H. Arai, J. Jpn. Oil Chem. So
20 (1971) 160.
[15] E.S.R. Gopal, in: P. Sherman (Ed.), Emulsion SciencAcademic Press, London, 1969.
[16] P. Walstra, in: P. Bechap (Ed.), Encyclopedia of Emulsi
Technology, vol. 1, Marcel Dekker, New York, 1983.
[17] T. Horiuchi, Research and Development for Nov
Emulsifiers and New Emulsifier Technology, CMC, Toky
1998.
[18] R.S. Schechter, A. Garcia, J. Lachaise, J. Colloid Inte
Sci. 204 (1998) 398.
[19] T. Shimanouchi, Analyses of IR Spectra, Nankodo, 198
pp. 100103.
[20] K. Kalyanasundaram, J.K. Thomas, J. Am. Chem. Soc.
(1986) 288.
[21] R.J. Hunter, Foundations in Colloid Science, vol.
Clarendon Press, Oxford, 1986, p. 601.[22] T. Sakai, K. Kamogawa, N. Momozawa, H. Sakai, M
Abe, submitted.
[23] A. Ito, K. Kamogawa, H. Sakai, Y. Kondou, N. Yoshin
M. Abe, J. Jpn. Oil. Chem. Soc. 45 (1996) 479.
[24] R.G. Alargova, I.I. Kochijashky, R. Zana, Langmuir
(1998) 1575.
[25] S. Zhang, J.F. Rusring, J. Colloid Interf. Sci. 182 (199
558.
[26] Y. Saito, Y. Kondo, M. Abe, T. Sato, J. Jpn. Oil Chem
Soc. 11 (1994) 325.
[27] R. Buscall, S.S. Davis, D.C. Potts, Colloid Polym. Sci. 2
(1979) 636.
[28] T. Shimanouchi, Analyses of IR Spectra, Nankodo, 198
p. 263.[29] I.M. Lifshitz, V.V. Slezov, J. Phys. Chem. Solids 19 (196
35.
[30] C. Wagner, Z. Electrochem. 65 (1961) 581.
[31] A.S. Kabalnov, Langmuir 10 (1994) 680.
[32] P. Taylor, Colloids Surf. A Physicochem. Eng. Aspects
(1995) 175.
[33] P. Taylor, Adv. Colloid Interf. Sci. 75 (1998) 107.