- Email: [email protected]
Double emulsions — scope, limitations and new achievements
COLLOIDS AND ELSEVIER
Colloids and Surfaces A: Physicochemical and Engineering Aspects 123-124 (1997) 233 246
A
SURFACES
Double emulsions -- scope, limitations and new achievements Nissim Garti CasaIi Institute of Applied Chemistry, The Hebrew University of Jerusalem, Jerusalem, 91904 Israel Received 7 May 1996; accepted 25 July 1996
Abstract Multiple emulsions are complex systems, termed "emulsions of emulsions", i.e. the droplets of the dispersed phase contain even smaller dispersed droplets themselves. Each dispersed globule in the double emulsion is separated from the aqueous phase by a layer of oil-phase compartments. Double emulsions have significant potential in many applications since, at least in theory, they can serve as an entrapping reservoir for active ingredients that can be released from the inner phase to the outer phase by a controlled and a sustained mechanism. Many of the potential applications are in pharmaceuticals. In practice, double emulsions are thermodynamically unstable systems with a strong tendency for coalescence, flocculation and creaming. Most double emulsions consist of relatively large droplets, cannot withstand storage regimes and have a strong tendency to release the entrapped matter in an uncontrolled manner. Much work has been devoted in the last decade in order to the improve the stability of the multiple emulsions and the control of the release rates of the addenda. The most recent achievements are: Use of polymeric emulsifiers to improve interface coverage and to better anchor into the dispersed phases; Reduce droplets size by improving methods of formation: improved understanding of the release mechanisms and use of various additives to control the release via the reverse micellar mechanism.
Keywords: Double emulsions; Multiple emulsions
1. Definitions, formation Multiple emulsions are complex systems, termed "emulsions of emulsions", i.e. the droplets of the dispersed phase contain even smaller dispersed droplets themselves. Each dispersed globule in the double emulsion forms a vesicular structure with single or multiple aqueous compartments separated from the aqueous phase by a layer of oil phase compartments [ 1 - 3 ] . Multiple emulsions were first described by Seifriz (see Ref. [-4]) in 1925, but it is only in the past 20 years that they have been studied in more detail.
The two major types of multiple emulsions are the water-oil-water (w/o/w) and oil-water-oil (o/w/o) double emulsions. A schematic representation of a w/o/w double emulsion droplet is shown in Fig. 1. Multiple emulsions have been prepared in two main modes, one-step emulsification and two-step emulsification. The most spontaneous preparation is based on forming w/o emulsion (the inner emulsion termed also the E1 emulsion) with a large excess of a relatively hydrophobic emulsifier and a small amount of hydrophilic emulsifier followed by heat-treating the emulsion until, at least in part,
0927-7757/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved PII S0927-7757 (96) 03809-5
234
N. Garti / Colloids Surjaces A: Physicochem. Eng. Aspects 123-124 (1997) 233 246
multiple (oil) drop interna! aqueousd
external aqueousphase
Prepar@l~on Methods: r
o
p
~
Two sleps :
'x~'~.N--~:,,:~:~-,?e.~-~m~.. ~ -,/:' STE P I
Fig. 1. Schematic representation of a w/o/w double emulsion droplet.
it inverts. At a proper temperature, and with the right hydrophilic-lipophilic balance (HLB) of the emulsifiers w/o/w emulsion can be found in the system. However, there is usually little chance of reproducing these "accidental" preparations. The most common and better controlled preparations of double emulsions are based on the two-step emulsification process by two sets of emulsifiers; a hydrophobic "Emulsifier I" (for the water-in-oil emulsion) and a hydrophilic - "Emulsifier II" (for the oil-in-water emulsion) (Fig. 2). The primary w/o emulsion is prepared under high shear conditions (ultrasonification, homogenization), while the secondary emulsification step is carried out without any severe mixing (an excess of mixing can rupture the drops resulting in a simple oil-in-water emulsion). The composition of the multiple emulsion is of significant importance since the different surfactants, along with the nature and concentration of the oil phase, will affect the stability of the double emulsion [ 1,3,5,6]. Much work has been done on the nature of oils in relation to the selected emulsifiers and their influence on the manufacturing conditions, as well as on the stability of the double emulsion [3]. The emulsifiers have been found to migrate from one interface to the other and alter the interfacial emulsifier organization. Thus, complicated calculations of apparent and weighted HLBs at each interface have been done [ 6 - 8 ] , and sophisticated combinations of emulsifiers have been considered (the HLB of the outer emulsion was found to be a weighted HLB of the contribution of the two
mix ,0, • homo~'~',~"~ genizalion
",~ee;.; ~"++J"
oil + Iipophitic surfactanl
?
W/O
emutsion
W/O emulsion ,~.~,,/~ f~J
///
STEP 2 hydrophilic surfaclant
-/
W/O/W
mulliple emulsion
Fig. 2. Schematicillustration of a two-step process in formation of a double emulsion.
types of emulsifiers) [9]. Mixed reversed micelles that can solubilize the water soluble active ingredients and transport them to the outer phase, have been detected in the lamellar oil phase (in w/o/w emulsions). Methods and means have been proposed to limit such solubilization. Ionic and non-ionic surfactants have been used for different applications, in accordance with their performance and health restrictions. However, it was well established that combinations of emulsifiers at the inner as well as at the outer phase have a beneficial effect on stability, and that the inner hydrophobic Emulsifier I must be used in great excess (10-30 wt.% of the inner emulsion) while the hydrophilic Emulsifier II must be used in low concentration (0.5-5 wt.%). In addition, parameters (Fig. 3) such as the oil phase volume and the nature of the entrapped materials in the inner phase have been discussed and optimized [7 9].
N. Garti /' Colloids" Surfaces A: Physicochem. Eng. Aspects 123 124 (1997) 233-246
235
100 ©
90
/
.,.....:..-,,-"¢-"
u3
n~
80 -
©
i
....7
!
70 @
o
/
,'
g
Tween 80 and POE-oley[ e t r . -
•
POE-lauryl ether and ester Tween 20
60
•t"
0 -~,
50 40
/
#
•
:
Xf X I
I
1
X
POE-nonylphenyl ether POE = polyoxyethylene
~
I,I
ITtl
1
10
t
I
T
t T~ll
100
Weight ratio of Span 80 to hydrophiZic emulsifier Fig. 3. Weight ratio of Span 80 in the oil phase to hydrophilic emulsifiers in the aqueous suspending fluid, affecting the formation of water-liquid paraffin-water emulsions owing to the two separated steps of emulsification.
2. Stability and transport mechanisms Many review articles have been written on the potential practical applications of the multiple emulsions and on the main problems associated with this technology, i.e. their inherent thermodynamic instability [ 10]. It was concluded that the classical double emulsions prepared with two sets of monomeric emulsifiers cannot provide long-term stability to the double emulsion. The w/o/w emulsions with relatively large droplets and short-term stability that are commonly obtained cannot be used in practice [ 11 ]. It was and still is very difficult to determine the relative stability of double emulsions. Several techniques have been applied including microscopic examination, counting the volume and the number of droplets immediately after preparation and after prolonged storage, freeze-etching techniques, viscosity measurements and quantitative estimation of addenda transported from the inner phase to the outer phase and vice versa. In addition, engulfment or shrinkage of the double emulsion in the presence of water migration in or out of the
droplets has been studied and interpreted in terms of stability (Fig. 4). Several possible mechanisms by which materials may be transported across the oil layer in the w/o/w systems have been proposed and discussed I1,4,6]. The most common is the molecular diffusion controlled mechanism of oil soluble matter. Diffusion through the oil phase can be controlled, in ionizable matter, by controlling its dissociation. The transport rates will be dependent on the nature of the entrapped material (including its dissociation constant), and on the oil, as well as on the pH of the aqueous phase [12]. At low pH values the entrapped matter (barbiturate as one example) would exist almost exclusively as the unionized form and so would be readily soluble in the oil phase. The drug could, therefore, pass easily across the oil layer to the internal aqueous phase containing a basic buffer, which would ionize the addenda that is now insoluble in the oil phase and would become trapped within the internal aqueous phase. This would then be carried out with the emulsion as it is, voided from the gastro-intestinal tract (Fig. 5). The drug transport was found to follow the first-order kinetics according to Fick's Law [ 12].
N. Garti / Colloids Surfaces A. Physicochem. Eng. Aspects 123-124 (1997) 233 246
236
t
External drop coalescence ~1t~ ~ primary ] ~w'- Ldispersion) Internal drop ~
growth ~
,
J
Primarydispersion
T
\
rl.idloss
~" Expulsi I" ~ o ' or n internal drops
Q EmRty primary continuous phase drop Fig. 4. Possible pathways for the transport of water and entrapped matter from the inner interface to the outer aqueous phase.
Ionized and ionizable compounds are not the only materials to be transported across the oil membrane. It has been demonstrated that both water molecules and non-electrolyte water-soluble matter can easily migrate through the oil membrane without affecting the double-emulsion stability [-13-15]. A mechanism based on "micellar transport" from one phase to the other has been described and determined. It was also demonstrated that the micellar transport is also diffusion controlled. One can alter the diffusion rates through the oil membrane by changing the nature of the oil, increasing its viscosity and adding various carriers. This suggests that the diffusion of the addenda through the oil is the rate determining step, and that the inner water phase does not have any effect on the determination of release rates. Kita, Matsumoto and Yonezawa [ 16] have suggested two possible mechanisms for the permeation of water and water soluble materials through the oil phase; the first being via "reverse micellar transport" (Fig. 6) and the second by "diffusion across a very thin lamella" of surfactant, formed where the oil layer is very thin (Fig. 7). Our detailed studies on the release of electrolytes from the inner aqueous phase to the outer aqueous phase in double emulsions stabilized by monomeric nonionic emulsifiers (Span and Tween) [13-15] have indicated that the osmotic pressure gradient between the two aqueous phases is a strong driving force for mutual migration of addenda and water from one phase through the other by both mechanisms. However, it was clearly demonstrated that even if the osmotic pressure of the two phases is equilibrated and no visual coalescence takes place, (neither of the inner phase droplets nor of the
vo'~ ~
Un i o n i s e d
{~queousphaseo-' ~o
dc~9
I -
I1'
*
A--r-
~-
P ,'Y--.AS
oil layer
inner oqu.phese
o
HA
Fig. 5. w/o/w "Liquid Membrane" system for removal of acidic drugs form an aqueous system.
Fig. 6. Schematic illustration of a model for micellar transport of water from the outer aqueous phase to the inner aqueous phase through the oil layer in w/o/w emulsion.
N. Garti / Colloids" Surfaces A." Physicochem. Eng. Aspects 123-124 (1997) 233-246
237
formation of relatively large double-emulsion droplets with very limited thermodynamic stability.
wa er
wal e r
3. New approaches to improve stability and release in double emulsions
o----o
[amella HYOROPHOBIC SURfXC-T^~
-----0
HYOROPf'([ L I C SURF^CT/kNT
----@
Fig. 7. Schematic illustration of a model for water transport through this lamella of the surfactant due to fluctuation in the thickness of the oil layer. outer phase droplets), electrolytes tend to be transported out, mostly through a reverse micellar mechanism controlled by the viscosity and the nature of the oil membrane. The mechanism is similar to the one described for release from polymeric matrices "slab into the sink" [16-18]. We adapted a modified diffusion controlled equation to explain the release results. It has been demonstrated that the release factor, B, B = 3/2[ 1 - ( 1 -
F ) 2/3 ] -
F
3Det
B ~ - -
ro2Co
where F is the release fraction of the marker soluble in the inner phase; De is effective diffusion coefficient; ro is radius of the droplets; t is time of release; Co is initial concentration of marker), can be plotted against 1~Co and t, with excellent correlation coefficients suggesting "diffusion-controlled release mechanism of reverse micellar transport" [14,15]. The "effective" as well as the "real" diffusion coefficient could therefore be calculated. Although the release mechanism was clarified to some extent, it is still very difficult to control the rate of release (to slow it down) mainly owing to the fact that the monomeric emulsifiers that have been used serve both as stabilizing moieties and as transport species, in addition, the fast exchange that those emulsifiers undergo between the two interfaces and the fact that shear should be avoided in the second emulsification stage lead to the
Fig. 4 demonstrates, in part, the different processes that can lead to the destabilization of the double emulsion and the processes involved in the release of various molecules (water or entrapped matter) from the inner phase to the outer aqueous phase. One can think of several approaches to overcome instability- and release-problems in double emulsions. Some of those ideas can be summarized as follows. The inner phase: (i) stabilizing the inner w/o emulsion by mechanically, or in the presence of better emulsifiers, reducing its droplet size; (ii) forming L2-microemulsions; (iii) preparing microspheres; (iv) increasing the viscosity of the inner water. The oil phase: (i) modifying the nature of the oil phase by increasing its viscosity or by adding carriers; (ii) adding complexing agents to the oil. The interfaces: (i) stabilizing the inner and/or the outer emulsion by using polymeric emulsifiers, macromolecular amphiphiles (proteins, polysaccharides) or colloidal solid particles to form strong and more rigid film at the interface. One can consider both naturally occurring macromolecules (gums, proteins) and synthetic grafted block copolymers with surface activities. It should be noted, although it is beyond the scope of this text, that polymerizable nonionic surfactants can form in-situ cross-linked membranes (after adsorption and carrying out a polymerization process). This concept was well documented and tested [19 21]. The polymeric complex that was formed in-situ was capable of withstanding extensive thinning (caused by osmotic driven influx of water) and the resulting swelling of the internal water droplets [22].
Naturally occurring macromolecular amphiphiles The stability of double emulsions can be improved (as explained above) by forming a poly-
238
N. Garti / Colloids SurJhces A: Physicochem. Eng. Aspects 123 124 (1997) 233 246
meric film or macromolecular complex across the oil/water interfaces. Omotosho et al. have suggested using macromolecules, and nonionic surfactants to form such stabilizing complexes. The film is formed through interfacial interaction between macromolecules such as albumin and nonionic surfactants [23,24]. Release rates of Methotrexate (MTX) encapsulated in the internal phase of w/o/w emulsions stabilized by a film, formed as a result of an interracial interaction between albumin and sorbitan monooleate (Span 80), were measured as functions of two formulation variables - - the oil phase and the secondary emulsifier composition (Fig. 8). The release rate was significantly affected by the nature of the oil phase and reflected the increasing internal droplet size of the emulsion. The release rate data conform with first order kinetics. Comparison of the effective permeability coefficients, calculated from the experimental apparent first-order rate constants, with the effective permeability coefficient of water in planar oil layers, containing non-ionic surfactants (determined by a microgravimetric method), supported the hypothesis of diffusion of M T X via loaded inverse micelles. Surfactants with high H L B values, used as the
secondary hydrophilic emulsifier, increased the release rates, primarily by increasing the rate of diffusion of MTX through the non-aqueous liquid membrane. Omotosho, Florence and co-workers have also reported use of other macromolecule complexes [23,24]. The formation of multiple w/o/w emulsions with improved stability owing to the formation of interfacial-complex film between acacia, gelatin, polyvinyl pyrrolidone (PVP) and sorbitan monooleate, have also been described [23,24] (Fig. 9). Multiple emulsions containing chloroquine phosphate in the internal phase that had been stored for two weeks, surprisingly showed a reduced rate of release of chloroquine phosphate as compared with freshly prepared emulsions, suggesting that the release from these systems occurs by the process of diffusion as opposed to the physical breakdown of emulsions [23]. Wasan and coworkers [25,26] have published a novel method for forming stable hemoglobin-oilin-water (Hb/o/w) multiple emulsion for use as an artificial red-cell substitute. A concentrated Hb solution was emulsified in oil to form microdroplets (with Pluronic F101 and Span 80), followed by dispersion of the primary emulsion into an outer
90 8o 70
•
IPM
60
@'
octadecar-
@
hexadec~
A
dodecane
4O 8
octane
30 20
10 0
'
I
1
r
I
2
r
1
3
r
I
I
4
I
5
T
I
6
T
I
7
T i m e (hours) Fig. 8. Releaseof MTX from multiple w/o/w emulsions. The emulsions contained 2.5% Span 80 and 0.2% BSA as primary emulsifiers, with MTX (1 mg m1-1) in the internal phase and the following oil phases: * - - octane; A - - dodecane; ,ll. - - hexadecane; ~ - octadecane; and • - - isopropylmyristate.
N. Garti / Colloids Surfaces A." Physicochem. Eng. Aspects 123 124 (1997) 233-246
239
80
60
A
A
40
•
contrc~
•
,o, 1
X
A2
V
c~
0
c2
2(?
0
0
2
4
6
Time (hours) Fig. 9. Profile of chloroquine phosphate release from w/o/w multiple emulsions. AI: freshly prepared PVP emulsions. Az: PVP emulsions stored for 2 weeks. BI: freshly prepared gelatin emulsions. B2: gelatin emulsions stored for 2 weeks, C~: emulsion prepared with gum acacia, C2: acacia emulsion stored for two weeks.
aqueous phase containing hydrophilic surfactants (Pluronic F68 or Tween 80). Addition of human serum albumin into both the inner and the outer phase along with added dextran into the outer phase, seemed to stabilize the emulsion. The average diameter of the prepared multiple emulsions after homogenization and filtration, was 2-3 pm with good hydrodynamic stability (sensitivity to shear). The formulation showed very small release of Hb from the primary emulsion to the outer aqueous phase, and good stability of the multiple emulsion during short-term storage. Oza and Frank [27,28] have suggested the use of colloidal microcrystalline cellulose (CMCC) and various monomeric surfactants to stabilize double emulsions via a mechanical stabilization mechanism. It was shown that the double emulsions were stable over a period of one month (monitored by microscopy). Slow release of certain drugs was achieved by gelling the oil phase. Recently, the author of this review and coworkers used BSA (bovine serum albumin) as a polymeric emulsifier added both to the inner and the outer interfaces, in absence and in presence of conventional monomeric emulsifiers (such as Span 80 and Tween 80) [29]. We tried to evaluate in more detail,
the mechanism of the release of an electrolyte, NaC1, from the inner phase to the outer phase in the presence of macromolecular emulsifiers. Since it was believed that the release mechanism is associated with a micellar transport which is diffusion controlled, attempts were made to obtain stable double-emulsions with relatively small droplets, and a minimum oil micellization capacity. Double emulsions were prepared with 10 wt.% Span 80, 0-0.5 wt.% BSA (as the inner emulsifying combination) and 2 wt.% NaC1 in the inner aqueous phase. The outer interface was stabilized with 5 wt.% of Span 80-Tween 80. It should be noted that after 25 h of aging, the size of the dispersed water-in-oil droplets hardly changed which indicates that the emulsion stability was maintained. The plot of the percent release of the entrapped matter, the electrolyte NaC1 versus wt.% of the inner BSA, reveals that BSA not only contributes to the stabilization of the emulsion, but also retards the release of NaC1 (Fig. 10). Its maximum effect was obtained at 0.2 wt.% BSA. It seems that BSA anchors to the interface, together with Span 80, and contributes both to the stability of the double emulsion and to the release rates.
240
N. Garti / Colloids Surfaces A. Physicochem. Eng. Aspects 123-124 (1997) 233-246
5 hours --O--
60 f
X 5O
se C/3
40
10 hours 15 hours 20 hours
<
---7--
25 hours
3O
(I)
20 10
t
1
o.~
T
1
I
0.2
I
03
T
1
0.4
r
I
05
BSA (wt %) i n irmer p h a s e Fig. 10. Percentage release with time of NaCI from a double emulsion prepared with 10 wt.% Span 80 and various BSA concentrations in the inner phase and 5 wt,% Span + Tween (1:9) in the outer aqueous phase.
In accordance with previous studies, it was suggested that the polymeric surfactant forms an interracial complex with the monomeric lipophilic surfactant. The complex is probably a thick, strongly gelled film that imparts elasticity and "resistance to rupture" of the inner droplets. The film improves the mechanical and steric stability of the double emulsions and slows the coalescence rates. In addition, it appears that it depresses the formation of reverse micelles in the oil phase and slows the transport of the electrolyte via the reverse micellar mechanism. However, the monomeric emulsifiers, if used alone to cover and stabilize the outer interface, do not prevent the coalescence of the outer droplets in the double emulsions. In fact, after one week of aging emulsions (stabilized with Tweens as external emulsifiers), a significant increase in droplet size distribution, as well as strong flocculation, was detected. Therefore, BSA was added (in addition to the monomeric Span 80-Tween 80) to the outer aqueous phase during the second step of emulsification. The Coulter-counter measurements (Fig. 11) clearly indicate that BSA, when present in the outer phase, contributes effectively to the stability
of the double emulsions. A double emulsion, prepared with BSA (in the outer interface), consisted of significantly smaller droplets than any other emulsions prepared without the BSA. Improved droplet size distribution was found for any emulsion prepared with BSA and any level of monomeric emulsifiers. The photomicrographs and the Coulter-counter measurements of the droplet size distribution after six weeks of aging, show practically no change in droplet size distribution, The release rates as a function of the BSA concentration in the outer phase (Emulsifier II) show also a minimum at 0.2 wt.% BSA and a slight increase in the release rates at higher BSA concentration. The Higuchi release model from solid polymeric matrix [17], as well as other models such as Garti's modification [ 14,15] for multiple emulsions (based on Fick's diffusion) were used for the double emulsions prepared with BSA in the inner phase. The B parameter (as described previously) was plotted against time (t) and versus the initial drug or electrolyte concentration (1/Co), in order to determine the effective diffusion coefficient (De). Figs. 12 and 13 are typical plots of the parameter
N. Garti / Colloids Surfaces A: Physicochem. Eng. Aspects 123-124 (1997) 233-246
0
10
20
30
40
50
60
70
80
90
r~
CURVE 1
~'
CURVE 2
"
CURVE 3
¢'
CURVE 4
241
100
D~ER(~m) Fig. 11. Droplet size distribution of four emulsions prepared: (1) with 10% Span 80+Tween 80 (9: 1) in the inner phase and 0.1 wt.% BSA in the outer aqueous phase; (2) with 10% Span+Tween 80 (9: 1) in the inner phase and without BSA in the outer phase; (3) with 10% Span 80 in the inner phase and 0.1% BSA in the outer phase; (4) with 10% Span 80 in the inner phase and without BSA in the outer phase. <2.(23 B 0.015
0.(]25 0.(]2
B
0.015 0.01
0.(]1 0.005 O
0.005
0
I
I
t
50
100
150
2o0
1 / C O (wt %)- 1 o f NaC1 '
I
1
2
3
4
5
Fig. 13. Same as Fig. 12 but plotting factor B against 1/Co of the NaCI concentration in the inner aqueous phase.
Time (hrs) Fig. 12. Factor B (see text) plotted against time of release for double emulsion stabilized with 10wt.% Span 80+0.2 wt.% BSA in the inner phase and 5 wt.% of Span 80/Tween 80 (9 : 1) in the outer phase (0.2% NaC1 in the inner phase).
B a g a i n s t I/Co o f e n t r a p p e d N a C 1 , a n d t h e a g i n g time of the emulsion. T h e De v a l u e s f o r e a c h B S A c o n c e n t r a t i o n c a l c u l a t e d f r o m p l o t B a g a i n s t l/Co a n d a g a i n s t t"(n=
N. Garti / Colloids Surfaces A: Physicochem. Eng. Aspects 123-124 (1997) 233 246
242
1), are linear and quite similar, indicating that the Higuchi model dominates the release mechanism. However, better correlation coefficients (r 2 = 0.998 to 1.000) will be obtained if the B parameter is plotted against t" where n (termed arbitrarily as the "diffusion order") varies from 0.5 to 3.0 as a function of the BSA concentration (see Table 1). Plots of parameter B versus t" for double emulsions stabilized with BSA in the outer phase, showed similar trends of linearity. The best correlation coefficients for linearity of the curves were obtained for t" where n was in the range of 0.5 1 for 0-0.5 wt.% BSA in the outerphase (W2) (see Table 1). The differences between the functionality and performance of the BSA present in the inner W1 or the outer W2 interface can be seen from the plot of the time exponent n (the "diffusion order") versus the BSA concentration (Fig. 14). The effective diffusion coefficients (Do) were calculated from the corrected Higuchi equation and plotted against the BSA concentrations in the inner and outer phases (Fig. 15). Significant differences in the performance of BSA, when added to the inner or to the outer interface, have been found. The outer BSA has only a limited retarding effect on the electrolytes' transport, limited to a concentration of 0-0.1 wt.%,
Table 1 Values of n ("diffusion order") obtained from plots of the B factor corresponding to the percentage of release of NaCI versus t" from double emulsions prepared with BSA in the inner phase (W~) and in the outer phase (Wz) (see text)
BSA in Wl (wt.%) 0 0.05 0.1 0.2 0.3 0.5
or W 2
(From t", for W l )
(Fromt", for W2)
0.5 1.25 2.0 2.0 2.0 3.0
0.5 0.5 0.5 0.5 1.0 1.0
nw~ is calculated from the best fit for linearity of the plot of B against t" for double emulsions stabilized with BSA added to the inner water phase (W1). nw2 is calculated from the best fit for linearity of the plot of factor B against t" of double emulsions stabilized with BSA added to the outer water phase (W2).
while the inner BSA (at 0.02 wt.% levels) has a pronounced slowing effect on the release of NaC1. It has been assumed that the parameter n is a reflection of the nature of the film that is formed on the interface. When n ~< 1 the film is rather thin and no significant viscoelastic gel (complex between the Span and BSA) is formed. However when n ~> 1 the film is viscous owing to the formation of a strong Span-BSA complex. It can be concluded that the internal w/o film (I1 interface) is more pronounced and stronger than the o/w (I2 interface) film. The I1 film develops as the BSA concentration increases and is well defined at 0.1-0.2 wt.% BSA. In contrast, the BSA in the outer phase (I2) does not contribute to the formation of a viscoelastic network with the hydrophilic combination of Span and Tween, and has only limited effect on the diffusion coefficient. Fig. 16 is a schematic illustration of the organization of the monomeric and polymeric emulsifiers onto the inner (I~) and the outer (I2) interfaces. Note that the BSA is co-adsorbed together with the monomeric emulsifier at the inner interface and serves only as a protective colloid at the outer interface. From a careful evaluation of the release and stability results it is possible to formulate an optimal double-emulsion consisting of Span 80 and BSA in the outer interface, and Span 80 and BSA in the inner interface. Similar results have been obtained in a more recent study carried out with emulsions stabilized be a combination of monomeric emulsifiers (Spans and Tweens) with polymeric, synthetic, well defined polymeric amphiphiles [30-33]. The inner interface was a blend of sorbitan esters and hydrophobic, ethoxylated, siliconic polymers. The external emulsifier was a blend of ethoxylated sorbitan esters with hydrophilic ethoxylated siliconic polymeric emulsifier. Much to our surprise, excellent double-emulsions were obtained (Fig. 17). The inner w/o emulsions were submicronal, while the outer double emulsions were shear resistant and allowed homogenization that yielded double emulsion droplets with narrow size distributions and in the range of 4-6/am (see Fig. 18). The double emulsions showed long-term stability (see Fig. 19) and slow release of the electrolytes or other addenda. It can be clearly
N. Garti / Colloids" Surfaces A." Physicochem. Eng. Aspects 123-124 (1997) 233-246
243
8
0 {D
2
[]
n(BSAin Wl)
•
n(BSAin W~I
0 ¢D
0.0
1
1
i
i
i
0.1
0.2
0.3
0.4
0.5
0.6
C O N C E N T R A T I O N o f B S A (w-t%) Fig. 14. Plot of exponent n of time (t") against BSA concentration in both inner ([3) and outer (O) interfaces. -16.0
]
-16.5 '
%, -17_0 O -17.5
EI
Iog(Do)(BSA it} W,.I
*
Log(De)(BSA ih ~,7~,
n
-18,0
0.0 0.~ 0.2 0 3 04 0.5 0.6 CONCENTRATION of BSA [wt%)
Fig. 15. Plot of log De (effective diffusion coefficient) of NaC1 against BSA concentration in the inner and the outer interfaces.
J ,/
Wt
'/
---' )G ©
~(
W2
--8SA
Fig. 16. Schematic illustration of the two interfaces of the double emulsion stabilized by a combination of monomeric and polymeric (BSA) emulsifiers. seen t h a t the p o l y m e r i c emulsifiers c o n t r i b u t e to the stability a n d to the c o n t r o l of the t r a n s p o r t of the a d d e n d a . T h e release rate d e p e n d s on the
n a t u r e of the m o n o m e r i c emulsifiers used at the inner interface a n d on the a m o u n t of the reversed micelles that are present in the oil phase. As the c o n c e n t r a t i o n of the Spans in the inner phase increases, so will the release rate increase. Evidences for the f o r m a t i o n of solubilized water a n d electrolytes have been d e m o n s t r a t e d . The polymeric emulsifiers, therefore, are an excellent tool to i m p r o v e b o t h stability a n d the t r a n s p o r t phenomenon. Recently, a t t e m p t s were m a d e to p r e p a r e d o u b l e emulsions stabilized with solid fat particles deposited at the internal a n d at the external interfaces. The results seem to be e n c o u r a g i n g [-34]. A n o t h e r a p p r o a c h is to emulsify L2 microemulsions [ 3 5 ] . T h e inner phase is a t h e r m o d y n a m i c a l l y stable m i c r o e m u l s i o n in which
N. Garti ,/Colloids SurJilces A." Physicochem, Eng. Aspects 123-124 (1997) 233 246
244
s0o] _~100% 7 0 0 ~ m~
0
ReLease
[]
]
>
80
I
~500 400
IS
~,
n
ETD
(a) 200
%% ~ ~ ? ~
O
Loo~. ~:]~ ;..~ .~ ~Abil . ~ EN~ *9 0 . ~
J 'O
L/ (b) Fig. 17. Photomicrographs of double emulsion stabilized with polymeric emulsifiers, after centrifugation. (a) magnification × 200; (b) magnification x 500.
4O
(5#m)
o~
v
Aqing time (davs2 o 0
2•
c 30
0
•
O o z~
~6 o 2O L
I
io °A
(I.) E I0
O
o
O
~o~ I _~ o~
!
15
30
4 Time
1
6 4 (clays)
,
I
io
12
Fig. 19. Plot of conductivity of the outer aqueous phase (reflecting the concentration of NaC1 in the outer phase (percentage release)) against time (days) in three sets of double emulsions. (1) All circles indicate use of Abil EM-90 as hydrophobic Emulsifier I. The lower curve (@) indicates the most hydrophilic P H M S P D M S U P E G emulsifier lI and the upper curve (©) in each set indicates the most hydrophobic P H M S P D M S U P E G with 52% substitution and 45 EO units. Each circle symbol represents a different polymeric emulsifier. (2) All triangles represent use of polyglycerol polyricinoleate (ETD) as Emulsifier 1 and the curves are arranged again with increasing hydrophobicity of emulsifier II. The lower curve (A) represents the most hydrophilic emulsifier and the upper curve in the set represents the most hydrophobic one (L.). (3) All squares represent the use of Span 80 as emulsifier 1 and the curves are arranged with increasing hydrophobicity of emulsifier II.
1
z~ 7 • 14 J, 21 o 30
O
o
(Siliconic) ,
Droplets Diameter (Nm) Fig. 18. Droplet size distribution of double emulsions with time.
addenda can be solubilized in the entrapped water. The microemulsion is thereafter emulsified into water containing hydrophilic emulsifiers and polymeric amphiphiles. The preliminary results are also quite encouraging.
4. Conclusions Double emulsions of w/o/w have many potential applications, but no real commercial product exists
N. Garti / Colloids Surfaces A." Physicochem. Eng. Aspects 123-124 (1997) 233 246
yet o n the market. The m a i n reason is the inherent instability of the p r e p a r a t i o n a n d the u n c o n t r o l l e d release of the a d d e n d a ( b o t h d u r i n g storage a n d within the time of application). The use of macromolecules to serve as steric stabilizers for b o t h the i n n e r and the outer interfaces have opened new options a n d possibilities. N a t u r a l l y occurring macromolecules such as selected proteins (BSA, HSA, gelatin) a n d hydrocolloids (gum arabic), have been used with great success to i m p r o v e the film f o r m a t i o n over the water a n d the oil phase (better anchoring, full coverage, thick layer, low desorption, no interfacial migration) together with m o n o m e r i c classical h y d r o p h o b i c a n d hydrophilic emulsifiers. The p o l y m e r - s u r f a c t a n t complex is an ideal interfacial barrier for diffusion-controlled t r a n s p o r t of b o t h h y d r o p h o b i c (non-ionized, lipid-like molecules) a n d hydrophilic (ionized molecules, hydrophilic organic molecules a n d electrolytes) substances. It has been s h o w n that both, t h e r m o d y n a m i c stability a n d e n t r a p m e n t , were significantly improved. F u r t h e r m o r e , the micellar t r a n s p o r t via reverse micelles was reduced, thus, m i g r a t i o n was very limited. The use of synthetic, well designed a n d characterized polymeric surfactants was very helpful in reducing leakage of addenda, i m p r o v i n g shear resistance and o b t a i n i n g small d o u b l e - e m u l s i o n s with excellent shelf-life a n d stability. For agricultural a n d industrial applications it seems to be an ideal formulation. Polymeric surfactants, in c o m b i n a t i o n with the c o n v e n t i o n a l small molecular-weight emulsifiers, are suggested as the future emulsifiers for double emulsions.
Acknowledgments I wish to t h a n k Dr. A b r a h a m Aserin a n d Mrs. A d r i a n a Yanai, for providing good ideas a n d helping to shape this manuscript.
References [1] S.S. Davis, J. Hadgraft and K.J. Palin, in P. Becher (Ed.), Encyclopedia of Emulsion Technology, Marcel Dekker, New York, Vol. 2, 159, (1985).
245
[2] A.T. Florence and D. Whitehill, Int. J. Pharm., 11 (1982) 277. [-3] A.T. Florence and D. Whitehill, J. Colloid Interface Sci., 79 (1981) 243. [-4] W. Seifriz, J. Phys. Chem., 29 (1980) 738. [-5] T.J. Lin, H. Kurihara and H. Ohta, J. Soc. Cosmet. Chem., 26 (1975) 121. [-6] S. Matsumoto and W.W. Kang, J. Dispersion Sci., Technol., 10 (1989) 455. [7] S. Matsumoto, Y. Kita and D. Yonezawa, J. Colloid Interface Sci., 57 (1976) 353. [8] A.T. Florence and D. Whitehill, J. Colloid Interface Sci., 79 (1981) 243. [9] M. Frenkel, R. Shwartz and N. Garti, J. Colloid Interface Sci., 94 (1983) 174. [10] N. Garti, M. Frenkel and R. Schwartz, J. Dispersion Sci. Technol., 4 (1983) 237. [11] A.F. Brodin and S.G. Frank, Acta Pharm. Suec., 15 (1978) 111. [12] C. Chang, G.C. Fuller, J.W. Frankenfeld and C.T. Rhodes, J. Pharm. Sci., 63 (1987). [, 13] S. Magdassi, M. Frenkel and N. Garti, J. Dispersion Sci. Technol., 5 (1984) 49. [14] S. Magdassi, M. Frenkel and N. Garti, Drug Ind. Pharm., 11 (1985) 791. [15] S. Magdassi and N. Garti, J. Control. Release, 3 (1986) 273. [16] Y. Kita, S. Matsumoto and D. Yonezawa, J. Colloid Interface Sci., 62 (1977) 87. [,17] T. Higuchi, J. Pharm. Sci., 52 (1963) 1145. [-18] T.K. Law, A.T. Florence and T.L. Whateley, J. Pharm. Pharmacol., 36 (1984) 50. [19] T.K. Law, T.L Whateley and A.T. Florence, Int. J. Pharm., 21 (1984) 277. [20] T.K. Law, T.L. Whateley and A.T. Florence, J. Control. Release, 3 (1986) 279. [,21] A.T. Florence, T.K. Law and T.L Whateley, J. Colloid Interface Sci., 107 (1985) 584. [22] J.A. Omotosho, T.K. Law, T.L Whateley and A.T. Florence, Colloids Surfaces, 20 (1986) 133. [23] J.A. Omotosho, T.K. Law, T.L. Whateley and A.T. Florence, J. Pharm. PharmacoL 38 (1986) 865. [24] J.A. Omotosho, T.K. Law, T.L. Whateley and A.T. Florence, J. Microencapsulation, 6 (1989) 183. [25] R.L. Beissinger,D.T. Wasan, L.R. Sehgal and A.L. Rosen, UK Patent GB 2,221,912, 1990. [26] S. Zheng, R.L. Beissinger and D.T. Wasan, J. Colloid Interface Sci., 144 (1991) 72. [-27] K.P. Oza and S.G. Frank, J. Dispersion Sci. Technol., 7 (1986) 543. [28] K.P. Oza and S.G. Frank, J. Dispersion Sci. TechnoL, 10 (1989) 187. [29] N. Garti, A. Aserin and Y. Cohen, J. Control. Release, 29 (1994) 41. [30] Y. Sela, S. Magdassi and N. Garti, J. Controlled Release, 33 (1995) 1.
246
N. Garti / Colloids Surfaces A: Physicochem. Eng. Aspects 123-124 (1997) 233-246
[31] Y. Sela, S. Magdassi and N. Garti, Colloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 143. [32] Y. Sela, S. Magdassi and N. Garti, Colloid Polym. Sci., 272 (1994) 684. [33] N. Garti, A. Aserin and Y. Meidani, Release of electrolytes from the inner water phase of double emulsions stabilized by lysozyme and nonionic surfactants, in preparation.
[34] H. Benjamin, Double emulsions stabilized by fat crystals, M. Sc. Thesis, Casali Institute of Applied Chemistry, The Hebrew University of Jerusalem, Israel, 1996. [35] N. Garti and H. Houminer, Emulsification of microemulsions [(w/o microemulsion) in water], in preparation.
Colloids and Surfaces A: Physicochemical and Engineering Aspects 123-124 (1997) 233 246
A
SURFACES
Double emulsions -- scope, limitations and new achievements Nissim Garti CasaIi Institute of Applied Chemistry, The Hebrew University of Jerusalem, Jerusalem, 91904 Israel Received 7 May 1996; accepted 25 July 1996
Abstract Multiple emulsions are complex systems, termed "emulsions of emulsions", i.e. the droplets of the dispersed phase contain even smaller dispersed droplets themselves. Each dispersed globule in the double emulsion is separated from the aqueous phase by a layer of oil-phase compartments. Double emulsions have significant potential in many applications since, at least in theory, they can serve as an entrapping reservoir for active ingredients that can be released from the inner phase to the outer phase by a controlled and a sustained mechanism. Many of the potential applications are in pharmaceuticals. In practice, double emulsions are thermodynamically unstable systems with a strong tendency for coalescence, flocculation and creaming. Most double emulsions consist of relatively large droplets, cannot withstand storage regimes and have a strong tendency to release the entrapped matter in an uncontrolled manner. Much work has been devoted in the last decade in order to the improve the stability of the multiple emulsions and the control of the release rates of the addenda. The most recent achievements are: Use of polymeric emulsifiers to improve interface coverage and to better anchor into the dispersed phases; Reduce droplets size by improving methods of formation: improved understanding of the release mechanisms and use of various additives to control the release via the reverse micellar mechanism.
Keywords: Double emulsions; Multiple emulsions
1. Definitions, formation Multiple emulsions are complex systems, termed "emulsions of emulsions", i.e. the droplets of the dispersed phase contain even smaller dispersed droplets themselves. Each dispersed globule in the double emulsion forms a vesicular structure with single or multiple aqueous compartments separated from the aqueous phase by a layer of oil phase compartments [ 1 - 3 ] . Multiple emulsions were first described by Seifriz (see Ref. [-4]) in 1925, but it is only in the past 20 years that they have been studied in more detail.
The two major types of multiple emulsions are the water-oil-water (w/o/w) and oil-water-oil (o/w/o) double emulsions. A schematic representation of a w/o/w double emulsion droplet is shown in Fig. 1. Multiple emulsions have been prepared in two main modes, one-step emulsification and two-step emulsification. The most spontaneous preparation is based on forming w/o emulsion (the inner emulsion termed also the E1 emulsion) with a large excess of a relatively hydrophobic emulsifier and a small amount of hydrophilic emulsifier followed by heat-treating the emulsion until, at least in part,
0927-7757/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved PII S0927-7757 (96) 03809-5
234
N. Garti / Colloids Surjaces A: Physicochem. Eng. Aspects 123-124 (1997) 233 246
multiple (oil) drop interna! aqueousd
external aqueousphase
Prepar@l~on Methods: r
o
p
~
Two sleps :
'x~'~.N--~:,,:~:~-,?e.~-~m~.. ~ -,/:' STE P I
Fig. 1. Schematic representation of a w/o/w double emulsion droplet.
it inverts. At a proper temperature, and with the right hydrophilic-lipophilic balance (HLB) of the emulsifiers w/o/w emulsion can be found in the system. However, there is usually little chance of reproducing these "accidental" preparations. The most common and better controlled preparations of double emulsions are based on the two-step emulsification process by two sets of emulsifiers; a hydrophobic "Emulsifier I" (for the water-in-oil emulsion) and a hydrophilic - "Emulsifier II" (for the oil-in-water emulsion) (Fig. 2). The primary w/o emulsion is prepared under high shear conditions (ultrasonification, homogenization), while the secondary emulsification step is carried out without any severe mixing (an excess of mixing can rupture the drops resulting in a simple oil-in-water emulsion). The composition of the multiple emulsion is of significant importance since the different surfactants, along with the nature and concentration of the oil phase, will affect the stability of the double emulsion [ 1,3,5,6]. Much work has been done on the nature of oils in relation to the selected emulsifiers and their influence on the manufacturing conditions, as well as on the stability of the double emulsion [3]. The emulsifiers have been found to migrate from one interface to the other and alter the interfacial emulsifier organization. Thus, complicated calculations of apparent and weighted HLBs at each interface have been done [ 6 - 8 ] , and sophisticated combinations of emulsifiers have been considered (the HLB of the outer emulsion was found to be a weighted HLB of the contribution of the two
mix ,0, • homo~'~',~"~ genizalion
",~ee;.; ~"++J"
oil + Iipophitic surfactanl
?
W/O
emutsion
W/O emulsion ,~.~,,/~ f~J
///
STEP 2 hydrophilic surfaclant
-/
W/O/W
mulliple emulsion
Fig. 2. Schematicillustration of a two-step process in formation of a double emulsion.
types of emulsifiers) [9]. Mixed reversed micelles that can solubilize the water soluble active ingredients and transport them to the outer phase, have been detected in the lamellar oil phase (in w/o/w emulsions). Methods and means have been proposed to limit such solubilization. Ionic and non-ionic surfactants have been used for different applications, in accordance with their performance and health restrictions. However, it was well established that combinations of emulsifiers at the inner as well as at the outer phase have a beneficial effect on stability, and that the inner hydrophobic Emulsifier I must be used in great excess (10-30 wt.% of the inner emulsion) while the hydrophilic Emulsifier II must be used in low concentration (0.5-5 wt.%). In addition, parameters (Fig. 3) such as the oil phase volume and the nature of the entrapped materials in the inner phase have been discussed and optimized [7 9].
N. Garti /' Colloids" Surfaces A: Physicochem. Eng. Aspects 123 124 (1997) 233-246
235
100 ©
90
/
.,.....:..-,,-"¢-"
u3
n~
80 -
©
i
....7
!
70 @
o
/
,'
g
Tween 80 and POE-oley[ e t r . -
•
POE-lauryl ether and ester Tween 20
60
•t"
0 -~,
50 40
/
#
•
:
Xf X I
I
1
X
POE-nonylphenyl ether POE = polyoxyethylene
~
I,I
ITtl
1
10
t
I
T
t T~ll
100
Weight ratio of Span 80 to hydrophiZic emulsifier Fig. 3. Weight ratio of Span 80 in the oil phase to hydrophilic emulsifiers in the aqueous suspending fluid, affecting the formation of water-liquid paraffin-water emulsions owing to the two separated steps of emulsification.
2. Stability and transport mechanisms Many review articles have been written on the potential practical applications of the multiple emulsions and on the main problems associated with this technology, i.e. their inherent thermodynamic instability [ 10]. It was concluded that the classical double emulsions prepared with two sets of monomeric emulsifiers cannot provide long-term stability to the double emulsion. The w/o/w emulsions with relatively large droplets and short-term stability that are commonly obtained cannot be used in practice [ 11 ]. It was and still is very difficult to determine the relative stability of double emulsions. Several techniques have been applied including microscopic examination, counting the volume and the number of droplets immediately after preparation and after prolonged storage, freeze-etching techniques, viscosity measurements and quantitative estimation of addenda transported from the inner phase to the outer phase and vice versa. In addition, engulfment or shrinkage of the double emulsion in the presence of water migration in or out of the
droplets has been studied and interpreted in terms of stability (Fig. 4). Several possible mechanisms by which materials may be transported across the oil layer in the w/o/w systems have been proposed and discussed I1,4,6]. The most common is the molecular diffusion controlled mechanism of oil soluble matter. Diffusion through the oil phase can be controlled, in ionizable matter, by controlling its dissociation. The transport rates will be dependent on the nature of the entrapped material (including its dissociation constant), and on the oil, as well as on the pH of the aqueous phase [12]. At low pH values the entrapped matter (barbiturate as one example) would exist almost exclusively as the unionized form and so would be readily soluble in the oil phase. The drug could, therefore, pass easily across the oil layer to the internal aqueous phase containing a basic buffer, which would ionize the addenda that is now insoluble in the oil phase and would become trapped within the internal aqueous phase. This would then be carried out with the emulsion as it is, voided from the gastro-intestinal tract (Fig. 5). The drug transport was found to follow the first-order kinetics according to Fick's Law [ 12].
N. Garti / Colloids Surfaces A. Physicochem. Eng. Aspects 123-124 (1997) 233 246
236
t
External drop coalescence ~1t~ ~ primary ] ~w'- Ldispersion) Internal drop ~
growth ~
,
J
Primarydispersion
T
\
rl.idloss
~" Expulsi I" ~ o ' or n internal drops
Q EmRty primary continuous phase drop Fig. 4. Possible pathways for the transport of water and entrapped matter from the inner interface to the outer aqueous phase.
Ionized and ionizable compounds are not the only materials to be transported across the oil membrane. It has been demonstrated that both water molecules and non-electrolyte water-soluble matter can easily migrate through the oil membrane without affecting the double-emulsion stability [-13-15]. A mechanism based on "micellar transport" from one phase to the other has been described and determined. It was also demonstrated that the micellar transport is also diffusion controlled. One can alter the diffusion rates through the oil membrane by changing the nature of the oil, increasing its viscosity and adding various carriers. This suggests that the diffusion of the addenda through the oil is the rate determining step, and that the inner water phase does not have any effect on the determination of release rates. Kita, Matsumoto and Yonezawa [ 16] have suggested two possible mechanisms for the permeation of water and water soluble materials through the oil phase; the first being via "reverse micellar transport" (Fig. 6) and the second by "diffusion across a very thin lamella" of surfactant, formed where the oil layer is very thin (Fig. 7). Our detailed studies on the release of electrolytes from the inner aqueous phase to the outer aqueous phase in double emulsions stabilized by monomeric nonionic emulsifiers (Span and Tween) [13-15] have indicated that the osmotic pressure gradient between the two aqueous phases is a strong driving force for mutual migration of addenda and water from one phase through the other by both mechanisms. However, it was clearly demonstrated that even if the osmotic pressure of the two phases is equilibrated and no visual coalescence takes place, (neither of the inner phase droplets nor of the
vo'~ ~
Un i o n i s e d
{~queousphaseo-' ~o
dc~9
I -
I1'
*
A--r-
~-
P ,'Y--.AS
oil layer
inner oqu.phese
o
HA
Fig. 5. w/o/w "Liquid Membrane" system for removal of acidic drugs form an aqueous system.
Fig. 6. Schematic illustration of a model for micellar transport of water from the outer aqueous phase to the inner aqueous phase through the oil layer in w/o/w emulsion.
N. Garti / Colloids" Surfaces A." Physicochem. Eng. Aspects 123-124 (1997) 233-246
237
formation of relatively large double-emulsion droplets with very limited thermodynamic stability.
wa er
wal e r
3. New approaches to improve stability and release in double emulsions
o----o
[amella HYOROPHOBIC SURfXC-T^~
-----0
HYOROPf'([ L I C SURF^CT/kNT
----@
Fig. 7. Schematic illustration of a model for water transport through this lamella of the surfactant due to fluctuation in the thickness of the oil layer. outer phase droplets), electrolytes tend to be transported out, mostly through a reverse micellar mechanism controlled by the viscosity and the nature of the oil membrane. The mechanism is similar to the one described for release from polymeric matrices "slab into the sink" [16-18]. We adapted a modified diffusion controlled equation to explain the release results. It has been demonstrated that the release factor, B, B = 3/2[ 1 - ( 1 -
F ) 2/3 ] -
F
3Det
B ~ - -
ro2Co
where F is the release fraction of the marker soluble in the inner phase; De is effective diffusion coefficient; ro is radius of the droplets; t is time of release; Co is initial concentration of marker), can be plotted against 1~Co and t, with excellent correlation coefficients suggesting "diffusion-controlled release mechanism of reverse micellar transport" [14,15]. The "effective" as well as the "real" diffusion coefficient could therefore be calculated. Although the release mechanism was clarified to some extent, it is still very difficult to control the rate of release (to slow it down) mainly owing to the fact that the monomeric emulsifiers that have been used serve both as stabilizing moieties and as transport species, in addition, the fast exchange that those emulsifiers undergo between the two interfaces and the fact that shear should be avoided in the second emulsification stage lead to the
Fig. 4 demonstrates, in part, the different processes that can lead to the destabilization of the double emulsion and the processes involved in the release of various molecules (water or entrapped matter) from the inner phase to the outer aqueous phase. One can think of several approaches to overcome instability- and release-problems in double emulsions. Some of those ideas can be summarized as follows. The inner phase: (i) stabilizing the inner w/o emulsion by mechanically, or in the presence of better emulsifiers, reducing its droplet size; (ii) forming L2-microemulsions; (iii) preparing microspheres; (iv) increasing the viscosity of the inner water. The oil phase: (i) modifying the nature of the oil phase by increasing its viscosity or by adding carriers; (ii) adding complexing agents to the oil. The interfaces: (i) stabilizing the inner and/or the outer emulsion by using polymeric emulsifiers, macromolecular amphiphiles (proteins, polysaccharides) or colloidal solid particles to form strong and more rigid film at the interface. One can consider both naturally occurring macromolecules (gums, proteins) and synthetic grafted block copolymers with surface activities. It should be noted, although it is beyond the scope of this text, that polymerizable nonionic surfactants can form in-situ cross-linked membranes (after adsorption and carrying out a polymerization process). This concept was well documented and tested [19 21]. The polymeric complex that was formed in-situ was capable of withstanding extensive thinning (caused by osmotic driven influx of water) and the resulting swelling of the internal water droplets [22].
Naturally occurring macromolecular amphiphiles The stability of double emulsions can be improved (as explained above) by forming a poly-
238
N. Garti / Colloids SurJhces A: Physicochem. Eng. Aspects 123 124 (1997) 233 246
meric film or macromolecular complex across the oil/water interfaces. Omotosho et al. have suggested using macromolecules, and nonionic surfactants to form such stabilizing complexes. The film is formed through interfacial interaction between macromolecules such as albumin and nonionic surfactants [23,24]. Release rates of Methotrexate (MTX) encapsulated in the internal phase of w/o/w emulsions stabilized by a film, formed as a result of an interracial interaction between albumin and sorbitan monooleate (Span 80), were measured as functions of two formulation variables - - the oil phase and the secondary emulsifier composition (Fig. 8). The release rate was significantly affected by the nature of the oil phase and reflected the increasing internal droplet size of the emulsion. The release rate data conform with first order kinetics. Comparison of the effective permeability coefficients, calculated from the experimental apparent first-order rate constants, with the effective permeability coefficient of water in planar oil layers, containing non-ionic surfactants (determined by a microgravimetric method), supported the hypothesis of diffusion of M T X via loaded inverse micelles. Surfactants with high H L B values, used as the
secondary hydrophilic emulsifier, increased the release rates, primarily by increasing the rate of diffusion of MTX through the non-aqueous liquid membrane. Omotosho, Florence and co-workers have also reported use of other macromolecule complexes [23,24]. The formation of multiple w/o/w emulsions with improved stability owing to the formation of interfacial-complex film between acacia, gelatin, polyvinyl pyrrolidone (PVP) and sorbitan monooleate, have also been described [23,24] (Fig. 9). Multiple emulsions containing chloroquine phosphate in the internal phase that had been stored for two weeks, surprisingly showed a reduced rate of release of chloroquine phosphate as compared with freshly prepared emulsions, suggesting that the release from these systems occurs by the process of diffusion as opposed to the physical breakdown of emulsions [23]. Wasan and coworkers [25,26] have published a novel method for forming stable hemoglobin-oilin-water (Hb/o/w) multiple emulsion for use as an artificial red-cell substitute. A concentrated Hb solution was emulsified in oil to form microdroplets (with Pluronic F101 and Span 80), followed by dispersion of the primary emulsion into an outer
90 8o 70
•
IPM
60
@'
octadecar-
@
hexadec~
A
dodecane
4O 8
octane
30 20
10 0
'
I
1
r
I
2
r
1
3
r
I
I
4
I
5
T
I
6
T
I
7
T i m e (hours) Fig. 8. Releaseof MTX from multiple w/o/w emulsions. The emulsions contained 2.5% Span 80 and 0.2% BSA as primary emulsifiers, with MTX (1 mg m1-1) in the internal phase and the following oil phases: * - - octane; A - - dodecane; ,ll. - - hexadecane; ~ - octadecane; and • - - isopropylmyristate.
N. Garti / Colloids Surfaces A." Physicochem. Eng. Aspects 123 124 (1997) 233-246
239
80
60
A
A
40
•
contrc~
•
,o, 1
X
A2
V
c~
0
c2
2(?
0
0
2
4
6
Time (hours) Fig. 9. Profile of chloroquine phosphate release from w/o/w multiple emulsions. AI: freshly prepared PVP emulsions. Az: PVP emulsions stored for 2 weeks. BI: freshly prepared gelatin emulsions. B2: gelatin emulsions stored for 2 weeks, C~: emulsion prepared with gum acacia, C2: acacia emulsion stored for two weeks.
aqueous phase containing hydrophilic surfactants (Pluronic F68 or Tween 80). Addition of human serum albumin into both the inner and the outer phase along with added dextran into the outer phase, seemed to stabilize the emulsion. The average diameter of the prepared multiple emulsions after homogenization and filtration, was 2-3 pm with good hydrodynamic stability (sensitivity to shear). The formulation showed very small release of Hb from the primary emulsion to the outer aqueous phase, and good stability of the multiple emulsion during short-term storage. Oza and Frank [27,28] have suggested the use of colloidal microcrystalline cellulose (CMCC) and various monomeric surfactants to stabilize double emulsions via a mechanical stabilization mechanism. It was shown that the double emulsions were stable over a period of one month (monitored by microscopy). Slow release of certain drugs was achieved by gelling the oil phase. Recently, the author of this review and coworkers used BSA (bovine serum albumin) as a polymeric emulsifier added both to the inner and the outer interfaces, in absence and in presence of conventional monomeric emulsifiers (such as Span 80 and Tween 80) [29]. We tried to evaluate in more detail,
the mechanism of the release of an electrolyte, NaC1, from the inner phase to the outer phase in the presence of macromolecular emulsifiers. Since it was believed that the release mechanism is associated with a micellar transport which is diffusion controlled, attempts were made to obtain stable double-emulsions with relatively small droplets, and a minimum oil micellization capacity. Double emulsions were prepared with 10 wt.% Span 80, 0-0.5 wt.% BSA (as the inner emulsifying combination) and 2 wt.% NaC1 in the inner aqueous phase. The outer interface was stabilized with 5 wt.% of Span 80-Tween 80. It should be noted that after 25 h of aging, the size of the dispersed water-in-oil droplets hardly changed which indicates that the emulsion stability was maintained. The plot of the percent release of the entrapped matter, the electrolyte NaC1 versus wt.% of the inner BSA, reveals that BSA not only contributes to the stabilization of the emulsion, but also retards the release of NaC1 (Fig. 10). Its maximum effect was obtained at 0.2 wt.% BSA. It seems that BSA anchors to the interface, together with Span 80, and contributes both to the stability of the double emulsion and to the release rates.
240
N. Garti / Colloids Surfaces A. Physicochem. Eng. Aspects 123-124 (1997) 233-246
5 hours --O--
60 f
X 5O
se C/3
40
10 hours 15 hours 20 hours
<
---7--
25 hours
3O
(I)
20 10
t
1
o.~
T
1
I
0.2
I
03
T
1
0.4
r
I
05
BSA (wt %) i n irmer p h a s e Fig. 10. Percentage release with time of NaCI from a double emulsion prepared with 10 wt.% Span 80 and various BSA concentrations in the inner phase and 5 wt,% Span + Tween (1:9) in the outer aqueous phase.
In accordance with previous studies, it was suggested that the polymeric surfactant forms an interracial complex with the monomeric lipophilic surfactant. The complex is probably a thick, strongly gelled film that imparts elasticity and "resistance to rupture" of the inner droplets. The film improves the mechanical and steric stability of the double emulsions and slows the coalescence rates. In addition, it appears that it depresses the formation of reverse micelles in the oil phase and slows the transport of the electrolyte via the reverse micellar mechanism. However, the monomeric emulsifiers, if used alone to cover and stabilize the outer interface, do not prevent the coalescence of the outer droplets in the double emulsions. In fact, after one week of aging emulsions (stabilized with Tweens as external emulsifiers), a significant increase in droplet size distribution, as well as strong flocculation, was detected. Therefore, BSA was added (in addition to the monomeric Span 80-Tween 80) to the outer aqueous phase during the second step of emulsification. The Coulter-counter measurements (Fig. 11) clearly indicate that BSA, when present in the outer phase, contributes effectively to the stability
of the double emulsions. A double emulsion, prepared with BSA (in the outer interface), consisted of significantly smaller droplets than any other emulsions prepared without the BSA. Improved droplet size distribution was found for any emulsion prepared with BSA and any level of monomeric emulsifiers. The photomicrographs and the Coulter-counter measurements of the droplet size distribution after six weeks of aging, show practically no change in droplet size distribution, The release rates as a function of the BSA concentration in the outer phase (Emulsifier II) show also a minimum at 0.2 wt.% BSA and a slight increase in the release rates at higher BSA concentration. The Higuchi release model from solid polymeric matrix [17], as well as other models such as Garti's modification [ 14,15] for multiple emulsions (based on Fick's diffusion) were used for the double emulsions prepared with BSA in the inner phase. The B parameter (as described previously) was plotted against time (t) and versus the initial drug or electrolyte concentration (1/Co), in order to determine the effective diffusion coefficient (De). Figs. 12 and 13 are typical plots of the parameter
N. Garti / Colloids Surfaces A: Physicochem. Eng. Aspects 123-124 (1997) 233-246
0
10
20
30
40
50
60
70
80
90
r~
CURVE 1
~'
CURVE 2
"
CURVE 3
¢'
CURVE 4
241
100
D~ER(~m) Fig. 11. Droplet size distribution of four emulsions prepared: (1) with 10% Span 80+Tween 80 (9: 1) in the inner phase and 0.1 wt.% BSA in the outer aqueous phase; (2) with 10% Span+Tween 80 (9: 1) in the inner phase and without BSA in the outer phase; (3) with 10% Span 80 in the inner phase and 0.1% BSA in the outer phase; (4) with 10% Span 80 in the inner phase and without BSA in the outer phase. <2.(23 B 0.015
0.(]25 0.(]2
B
0.015 0.01
0.(]1 0.005 O
0.005
0
I
I
t
50
100
150
2o0
1 / C O (wt %)- 1 o f NaC1 '
I
1
2
3
4
5
Fig. 13. Same as Fig. 12 but plotting factor B against 1/Co of the NaCI concentration in the inner aqueous phase.
Time (hrs) Fig. 12. Factor B (see text) plotted against time of release for double emulsion stabilized with 10wt.% Span 80+0.2 wt.% BSA in the inner phase and 5 wt.% of Span 80/Tween 80 (9 : 1) in the outer phase (0.2% NaC1 in the inner phase).
B a g a i n s t I/Co o f e n t r a p p e d N a C 1 , a n d t h e a g i n g time of the emulsion. T h e De v a l u e s f o r e a c h B S A c o n c e n t r a t i o n c a l c u l a t e d f r o m p l o t B a g a i n s t l/Co a n d a g a i n s t t"(n=
N. Garti / Colloids Surfaces A: Physicochem. Eng. Aspects 123-124 (1997) 233 246
242
1), are linear and quite similar, indicating that the Higuchi model dominates the release mechanism. However, better correlation coefficients (r 2 = 0.998 to 1.000) will be obtained if the B parameter is plotted against t" where n (termed arbitrarily as the "diffusion order") varies from 0.5 to 3.0 as a function of the BSA concentration (see Table 1). Plots of parameter B versus t" for double emulsions stabilized with BSA in the outer phase, showed similar trends of linearity. The best correlation coefficients for linearity of the curves were obtained for t" where n was in the range of 0.5 1 for 0-0.5 wt.% BSA in the outerphase (W2) (see Table 1). The differences between the functionality and performance of the BSA present in the inner W1 or the outer W2 interface can be seen from the plot of the time exponent n (the "diffusion order") versus the BSA concentration (Fig. 14). The effective diffusion coefficients (Do) were calculated from the corrected Higuchi equation and plotted against the BSA concentrations in the inner and outer phases (Fig. 15). Significant differences in the performance of BSA, when added to the inner or to the outer interface, have been found. The outer BSA has only a limited retarding effect on the electrolytes' transport, limited to a concentration of 0-0.1 wt.%,
Table 1 Values of n ("diffusion order") obtained from plots of the B factor corresponding to the percentage of release of NaCI versus t" from double emulsions prepared with BSA in the inner phase (W~) and in the outer phase (Wz) (see text)
BSA in Wl (wt.%) 0 0.05 0.1 0.2 0.3 0.5
or W 2
(From t", for W l )
(Fromt", for W2)
0.5 1.25 2.0 2.0 2.0 3.0
0.5 0.5 0.5 0.5 1.0 1.0
nw~ is calculated from the best fit for linearity of the plot of B against t" for double emulsions stabilized with BSA added to the inner water phase (W1). nw2 is calculated from the best fit for linearity of the plot of factor B against t" of double emulsions stabilized with BSA added to the outer water phase (W2).
while the inner BSA (at 0.02 wt.% levels) has a pronounced slowing effect on the release of NaC1. It has been assumed that the parameter n is a reflection of the nature of the film that is formed on the interface. When n ~< 1 the film is rather thin and no significant viscoelastic gel (complex between the Span and BSA) is formed. However when n ~> 1 the film is viscous owing to the formation of a strong Span-BSA complex. It can be concluded that the internal w/o film (I1 interface) is more pronounced and stronger than the o/w (I2 interface) film. The I1 film develops as the BSA concentration increases and is well defined at 0.1-0.2 wt.% BSA. In contrast, the BSA in the outer phase (I2) does not contribute to the formation of a viscoelastic network with the hydrophilic combination of Span and Tween, and has only limited effect on the diffusion coefficient. Fig. 16 is a schematic illustration of the organization of the monomeric and polymeric emulsifiers onto the inner (I~) and the outer (I2) interfaces. Note that the BSA is co-adsorbed together with the monomeric emulsifier at the inner interface and serves only as a protective colloid at the outer interface. From a careful evaluation of the release and stability results it is possible to formulate an optimal double-emulsion consisting of Span 80 and BSA in the outer interface, and Span 80 and BSA in the inner interface. Similar results have been obtained in a more recent study carried out with emulsions stabilized be a combination of monomeric emulsifiers (Spans and Tweens) with polymeric, synthetic, well defined polymeric amphiphiles [30-33]. The inner interface was a blend of sorbitan esters and hydrophobic, ethoxylated, siliconic polymers. The external emulsifier was a blend of ethoxylated sorbitan esters with hydrophilic ethoxylated siliconic polymeric emulsifier. Much to our surprise, excellent double-emulsions were obtained (Fig. 17). The inner w/o emulsions were submicronal, while the outer double emulsions were shear resistant and allowed homogenization that yielded double emulsion droplets with narrow size distributions and in the range of 4-6/am (see Fig. 18). The double emulsions showed long-term stability (see Fig. 19) and slow release of the electrolytes or other addenda. It can be clearly
N. Garti / Colloids" Surfaces A." Physicochem. Eng. Aspects 123-124 (1997) 233-246
243
8
0 {D
2
[]
n(BSAin Wl)
•
n(BSAin W~I
0 ¢D
0.0
1
1
i
i
i
0.1
0.2
0.3
0.4
0.5
0.6
C O N C E N T R A T I O N o f B S A (w-t%) Fig. 14. Plot of exponent n of time (t") against BSA concentration in both inner ([3) and outer (O) interfaces. -16.0
]
-16.5 '
%, -17_0 O -17.5
EI
Iog(Do)(BSA it} W,.I
*
Log(De)(BSA ih ~,7~,
n
-18,0
0.0 0.~ 0.2 0 3 04 0.5 0.6 CONCENTRATION of BSA [wt%)
Fig. 15. Plot of log De (effective diffusion coefficient) of NaC1 against BSA concentration in the inner and the outer interfaces.
J ,/
Wt
'/
---' )G ©
~(
W2
--8SA
Fig. 16. Schematic illustration of the two interfaces of the double emulsion stabilized by a combination of monomeric and polymeric (BSA) emulsifiers. seen t h a t the p o l y m e r i c emulsifiers c o n t r i b u t e to the stability a n d to the c o n t r o l of the t r a n s p o r t of the a d d e n d a . T h e release rate d e p e n d s on the
n a t u r e of the m o n o m e r i c emulsifiers used at the inner interface a n d on the a m o u n t of the reversed micelles that are present in the oil phase. As the c o n c e n t r a t i o n of the Spans in the inner phase increases, so will the release rate increase. Evidences for the f o r m a t i o n of solubilized water a n d electrolytes have been d e m o n s t r a t e d . The polymeric emulsifiers, therefore, are an excellent tool to i m p r o v e b o t h stability a n d the t r a n s p o r t phenomenon. Recently, a t t e m p t s were m a d e to p r e p a r e d o u b l e emulsions stabilized with solid fat particles deposited at the internal a n d at the external interfaces. The results seem to be e n c o u r a g i n g [-34]. A n o t h e r a p p r o a c h is to emulsify L2 microemulsions [ 3 5 ] . T h e inner phase is a t h e r m o d y n a m i c a l l y stable m i c r o e m u l s i o n in which
N. Garti ,/Colloids SurJilces A." Physicochem, Eng. Aspects 123-124 (1997) 233 246
244
s0o] _~100% 7 0 0 ~ m~
0
ReLease
[]
]
>
80
I
~500 400
IS
~,
n
ETD
(a) 200
%% ~ ~ ? ~
O
Loo~. ~:]~ ;..~ .~ ~Abil . ~ EN~ *9 0 . ~
J 'O
L/ (b) Fig. 17. Photomicrographs of double emulsion stabilized with polymeric emulsifiers, after centrifugation. (a) magnification × 200; (b) magnification x 500.
4O
(5#m)
o~
v
Aqing time (davs2 o 0
2•
c 30
0
•
O o z~
~6 o 2O L
I
io °A
(I.) E I0
O
o
O
~o~ I _~ o~
!
15
30
4 Time
1
6 4 (clays)
,
I
io
12
Fig. 19. Plot of conductivity of the outer aqueous phase (reflecting the concentration of NaC1 in the outer phase (percentage release)) against time (days) in three sets of double emulsions. (1) All circles indicate use of Abil EM-90 as hydrophobic Emulsifier I. The lower curve (@) indicates the most hydrophilic P H M S P D M S U P E G emulsifier lI and the upper curve (©) in each set indicates the most hydrophobic P H M S P D M S U P E G with 52% substitution and 45 EO units. Each circle symbol represents a different polymeric emulsifier. (2) All triangles represent use of polyglycerol polyricinoleate (ETD) as Emulsifier 1 and the curves are arranged again with increasing hydrophobicity of emulsifier II. The lower curve (A) represents the most hydrophilic emulsifier and the upper curve in the set represents the most hydrophobic one (L.). (3) All squares represent the use of Span 80 as emulsifier 1 and the curves are arranged with increasing hydrophobicity of emulsifier II.
1
z~ 7 • 14 J, 21 o 30
O
o
(Siliconic) ,
Droplets Diameter (Nm) Fig. 18. Droplet size distribution of double emulsions with time.
addenda can be solubilized in the entrapped water. The microemulsion is thereafter emulsified into water containing hydrophilic emulsifiers and polymeric amphiphiles. The preliminary results are also quite encouraging.
4. Conclusions Double emulsions of w/o/w have many potential applications, but no real commercial product exists
N. Garti / Colloids Surfaces A." Physicochem. Eng. Aspects 123-124 (1997) 233 246
yet o n the market. The m a i n reason is the inherent instability of the p r e p a r a t i o n a n d the u n c o n t r o l l e d release of the a d d e n d a ( b o t h d u r i n g storage a n d within the time of application). The use of macromolecules to serve as steric stabilizers for b o t h the i n n e r and the outer interfaces have opened new options a n d possibilities. N a t u r a l l y occurring macromolecules such as selected proteins (BSA, HSA, gelatin) a n d hydrocolloids (gum arabic), have been used with great success to i m p r o v e the film f o r m a t i o n over the water a n d the oil phase (better anchoring, full coverage, thick layer, low desorption, no interfacial migration) together with m o n o m e r i c classical h y d r o p h o b i c a n d hydrophilic emulsifiers. The p o l y m e r - s u r f a c t a n t complex is an ideal interfacial barrier for diffusion-controlled t r a n s p o r t of b o t h h y d r o p h o b i c (non-ionized, lipid-like molecules) a n d hydrophilic (ionized molecules, hydrophilic organic molecules a n d electrolytes) substances. It has been s h o w n that both, t h e r m o d y n a m i c stability a n d e n t r a p m e n t , were significantly improved. F u r t h e r m o r e , the micellar t r a n s p o r t via reverse micelles was reduced, thus, m i g r a t i o n was very limited. The use of synthetic, well designed a n d characterized polymeric surfactants was very helpful in reducing leakage of addenda, i m p r o v i n g shear resistance and o b t a i n i n g small d o u b l e - e m u l s i o n s with excellent shelf-life a n d stability. For agricultural a n d industrial applications it seems to be an ideal formulation. Polymeric surfactants, in c o m b i n a t i o n with the c o n v e n t i o n a l small molecular-weight emulsifiers, are suggested as the future emulsifiers for double emulsions.
Acknowledgments I wish to t h a n k Dr. A b r a h a m Aserin a n d Mrs. A d r i a n a Yanai, for providing good ideas a n d helping to shape this manuscript.
References [1] S.S. Davis, J. Hadgraft and K.J. Palin, in P. Becher (Ed.), Encyclopedia of Emulsion Technology, Marcel Dekker, New York, Vol. 2, 159, (1985).
245
[2] A.T. Florence and D. Whitehill, Int. J. Pharm., 11 (1982) 277. [-3] A.T. Florence and D. Whitehill, J. Colloid Interface Sci., 79 (1981) 243. [-4] W. Seifriz, J. Phys. Chem., 29 (1980) 738. [-5] T.J. Lin, H. Kurihara and H. Ohta, J. Soc. Cosmet. Chem., 26 (1975) 121. [-6] S. Matsumoto and W.W. Kang, J. Dispersion Sci., Technol., 10 (1989) 455. [7] S. Matsumoto, Y. Kita and D. Yonezawa, J. Colloid Interface Sci., 57 (1976) 353. [8] A.T. Florence and D. Whitehill, J. Colloid Interface Sci., 79 (1981) 243. [9] M. Frenkel, R. Shwartz and N. Garti, J. Colloid Interface Sci., 94 (1983) 174. [10] N. Garti, M. Frenkel and R. Schwartz, J. Dispersion Sci. Technol., 4 (1983) 237. [11] A.F. Brodin and S.G. Frank, Acta Pharm. Suec., 15 (1978) 111. [12] C. Chang, G.C. Fuller, J.W. Frankenfeld and C.T. Rhodes, J. Pharm. Sci., 63 (1987). [, 13] S. Magdassi, M. Frenkel and N. Garti, J. Dispersion Sci. Technol., 5 (1984) 49. [14] S. Magdassi, M. Frenkel and N. Garti, Drug Ind. Pharm., 11 (1985) 791. [15] S. Magdassi and N. Garti, J. Control. Release, 3 (1986) 273. [16] Y. Kita, S. Matsumoto and D. Yonezawa, J. Colloid Interface Sci., 62 (1977) 87. [,17] T. Higuchi, J. Pharm. Sci., 52 (1963) 1145. [-18] T.K. Law, A.T. Florence and T.L. Whateley, J. Pharm. Pharmacol., 36 (1984) 50. [19] T.K. Law, T.L Whateley and A.T. Florence, Int. J. Pharm., 21 (1984) 277. [20] T.K. Law, T.L. Whateley and A.T. Florence, J. Control. Release, 3 (1986) 279. [,21] A.T. Florence, T.K. Law and T.L Whateley, J. Colloid Interface Sci., 107 (1985) 584. [22] J.A. Omotosho, T.K. Law, T.L Whateley and A.T. Florence, Colloids Surfaces, 20 (1986) 133. [23] J.A. Omotosho, T.K. Law, T.L. Whateley and A.T. Florence, J. Pharm. PharmacoL 38 (1986) 865. [24] J.A. Omotosho, T.K. Law, T.L. Whateley and A.T. Florence, J. Microencapsulation, 6 (1989) 183. [25] R.L. Beissinger,D.T. Wasan, L.R. Sehgal and A.L. Rosen, UK Patent GB 2,221,912, 1990. [26] S. Zheng, R.L. Beissinger and D.T. Wasan, J. Colloid Interface Sci., 144 (1991) 72. [-27] K.P. Oza and S.G. Frank, J. Dispersion Sci. Technol., 7 (1986) 543. [28] K.P. Oza and S.G. Frank, J. Dispersion Sci. TechnoL, 10 (1989) 187. [29] N. Garti, A. Aserin and Y. Cohen, J. Control. Release, 29 (1994) 41. [30] Y. Sela, S. Magdassi and N. Garti, J. Controlled Release, 33 (1995) 1.
246
N. Garti / Colloids Surfaces A: Physicochem. Eng. Aspects 123-124 (1997) 233-246
[31] Y. Sela, S. Magdassi and N. Garti, Colloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 143. [32] Y. Sela, S. Magdassi and N. Garti, Colloid Polym. Sci., 272 (1994) 684. [33] N. Garti, A. Aserin and Y. Meidani, Release of electrolytes from the inner water phase of double emulsions stabilized by lysozyme and nonionic surfactants, in preparation.
[34] H. Benjamin, Double emulsions stabilized by fat crystals, M. Sc. Thesis, Casali Institute of Applied Chemistry, The Hebrew University of Jerusalem, Israel, 1996. [35] N. Garti and H. Houminer, Emulsification of microemulsions [(w/o microemulsion) in water], in preparation.