4808
Ind. Eng. Chem. Res. 2001, 40, 4808-4814
Influence of the Stirrer Initial Position on Emulsion Morphology. Making Use of the Local Water-to-Oil Ratio Concept for Formulation Engineering Purpose Serge E. Salager, Eric C. Tyrode, Marı´a-Teresa Celis, and Jean-Louis Salager* Laboratory FIRP, Ingenierı´a Quı´mica, Universidad de Los Andes, Me´ rida, Venezuela
The initial location of the stirrer in the emulsification vessel can induce the resulting emulsion type. The interpretation in terms of mixing phenomena leads to the use of the local water-to-oil ratio (WOR) concept in the formulation-composition map. The know-how associated with this phenomenology allows us to interpret in a straightforward way the kind of complex procedures commonly employed in emulsion manufacturing, particularly those associated with inversion and multiple emulsion attainment. Introduction In a house-made mayonnaise preparation, oil is progressively incorporated as droplets into the egg yolk, which is actually an aqueous solution containing oil-inwater (O/W) emulsion stabilizing agents. More and more oil is added while maintaining a gentle stirring, often provided by a spoon or a hand beater. The final product is an O/W emulsion that could contain an extremely high proportion of oil internal phase, says 90% or more. On the other hand, if this same amount of oil were initially mixed as a whole with the egg yolk, even under extremely energetic stirring, there would be no way to attain a mayonnaise, but a water-in-oil (W/O) egg yolk dispersion in oil instead. The preparation of mayonnaise is a well-known example of the hysteresis exhibited by emulsions, whose morphology depends on the way they are manufactured. This kind of behavior, which is related to the concept of “memory”, is not uncommon in nature and can be modeled by catastrophe theory in some elegant way.1-5 Emulsion morphology occurrence and its change from one type to the other, i.e., emulsion inversion, has regained interest lately and several research groups have dedicated their efforts to related topics. A summary of the general understanding and associated know-how may be found in a recent review.6 Coming back to the case of mayonnaise manufacturing, it is worth remarking that some kitchen appliances offer a possibility to go around the 90% oil proportion handicap, and to attain a O/W emulsion type by mixing a large amount of oil with the egg yolk (almost) at once, according to a principle proposed quite a long time ago.7 These devices are made of a turbine blender located at the end of a stem, which is placed at the bottom of a beaker-like vessel. Just above the moving impeller, there is some kind of cap, whose diameter is close to the vessel diameter, and which is outfitted with some foot that keeps the turbine at some distance, 1 in. or so, from the bottom of the container. The role of the cap is not only to protect the user from the rotating blade and to avoid spattering, but also to impede any direct * Corresponding author. Lab. FIRP, Ingenierı´a Quı´mica, Universidad de Los Andes, 5101 Me´rida, Venezuela. Phone: ++58-274-240-2954. Fax: ++58-274-240-2957. E-mail:
[email protected].
fluid convection in the axial direction. With such a restrain, the stirrer starts blending what happens to be confined below the cap, i.e., the egg yolk and a very small amount of oil, no matter the quantity of oil present in the rest of the container. After a few seconds of stirring, the stem is risen slightly, so that some oil is sucked in the emulsification zone, then it is lowered for more blending, then risen a little again and so on until the entire amount of oil becomes incorporated into the emulsion. Actually, this process exactly produces the same transient situation that the drop by drop addition of oil in a bowl, i.e., a gradual change in water-to-oil ratio in the fluid region concerned by the stirring process. However the whole system composition is unchanged. This is why it is quite pertinent to introduce the concept of instant or local water-to-oil ratio (WOR)8 to carry out the analysis of emulsion morphology occurrence and related inversion phenomena. This concept of local composition is also quite convenient to describe the emulsification in a large vessel in which the stirring device is too small to mix the whole system at once. Since an emulsion can retain the memory of its morphology, what happens during the very first instants of the emulsification could determine the issue. However, there is essentially no way to probe, and even less to scrutinize, the onset on emulsification in most practical cases, particularly in turbulent laboratory blenders, because everything is settled in a glance. To avoid this indetermination, the present study uses a stirring device which is purposely selected to be a very inefficient blender, so that it is possible to analyze the progress of the mixing process over some measurable time scale, e.g., at least a few seconds or minutes. The impeller device is a rotating disk, which can be placed in a horizontal position at, above, or below the original interface, and whose mixing action is linked with the overall fluid motion pattern created in the vessel, since it does not drive any direct axial motion. The purpose of this paper is to report and interpret the influence of the initial position of such a stirrer on the outcome of the emulsification procedure. Basic Concepts The influence of the stirrer position in the emulsification procedure, and its coupled effect with the formula-
10.1021/ie010196r CCC: $20.00 © 2001 American Chemical Society Published on Web 10/04/2001
Ind. Eng. Chem. Res., Vol. 40, No. 22, 2001 4809
Figure 1. Formulation-composition map.
tion and composition variables in the emulsion final properties, may be easily explained using the so-called formulation-composition map, which was first introduced almost 20 years ago,9 and its now a widespread tool10-13 used for the understanding of the different phenomena involved in the emulsion making. The formulation-composition map has not only been used to plot the emulsion type but also to display other important emulsion properties such as drop size, stability, and viscosity,14 just to mention a few. Figure 1a shows a formulation composition diagram at a constant surfactant concentration, where the formulation is expressed along the ordinate as a generalized parameter which indicates the hydrophiliclipophilic balance of the system, such as SAD (“surfactant affinity difference”) or the more recently introduced HLD (“hydrophilic lipophilic deviation”, i.e., the SAD value related to a three-phase behavior reference).15 Along the abscissa is the composition variable, which is generally taken as the water/oil content. The generalized formulation parameters SAD or HLD main interest is to compress all the formulation variables into a single one. The hydrophilic-lipophilic deviation to optimum formulation for three phase behavior (often referred to as Winsor III case) is defined by
HLD ) SAD/RT - ln(Co/Cw)WIII
(1)
where SAD is the molecular free energy of transfer of a surfactant molecule from oil to water, that is, ∆µofw, and (Co/Cw)WIII is the partition coefficient of the surfactant between oil and water at optimum formulation (reference case). It is worth noting that for ionic surfactant systems the partition coefficient is essentially unity at optimum formulation, thus making HLD equal to SAD/RT. Whether HLD is respectively positive or negative, the dominant affinity of the surfactant is respectively for the oil or water phase. At HLD ) 0 the equilibrium of affinity of the surfactant for both water and oil phases is attained. An up to date review on the generalized formulation concept can be found elsewhere.16 For now it is enough to mention that HLD varies with salinity, oil nature, and alcohol according to
HLD ) ln S - kACN + F(A) ...
(2)
where S is the salinity of the aqueous phase (in wt % NaCl), ACN is the number of carbon atoms in the alkane molecule (or its equivalent EACN if it is not a pure alkane), and F(A) is a function which increases as the n-pentanol content increases. Coming back to Figure 1a, the “standard” inversion line (black bold curve) represents the frontier between the different emulsion type occurrence, i.e., O/W and
W/O, which are found when an equilibrated SOW system is intensively stirred following a standard procedure. The horizontal branch and the two vertical branches of the standard inversion line are known as the transitional inversion frontier and the catastrophic inversion boundaries, respectively. The formulationcomposition map is divided into six zones (A+, A-, B+, B-, C+, and C-). The “+” and “-“ labels are directly related to the formulation parameter’s sign, that is, to a positive and negative HLD value, respectively. On the other hand, zones labeled as “A”, “B”, and “C” differ in their water content. Typically the A zones extend from 30 to 70% of water (or oil) content. Both A+ and Aregions and adjacent B+ and C- regions are so-called “normal” regions, because in these zones the emulsion type corresponds to the preferred morphology according to the Bancroft’s rule.17 In the two remaining regions, B- and C+, the preferred continuous phase content is too low for the “normal” emulsion to be formed and “abnormal” emulsions are formed instead. This means that in abnormal emulsions water is the external phase at HLD > 0 (C+ region) and oil the external phase at HLD < 0 (B- region). Abnormal emulsions are quite unstable, since their interfacial curvature is opposite to the one naturally induced by the surfactant. A general presentation of the formulation-composition map, with the emulsion associated properties can be found in a recent review.18 When an emulsion initially formulated in the A+ or A- zone, is subjected to a formulation change (e.g., salinity when using an ionic surfactant) at a constant water content and stirring conditions, its representative point in the formulation-composition map moves along a vertical line according to a continuous or lump wise motion. If the formulation is kept changing, the emulsion finally reaches the transitional inversion line where a morphology swap takes place due to the change in surfactant affinity. That is, the dispersed phase becomes the external phase and vice versa (vertical arrows in Figure 1b). On the other hand, if the formulation is kept constant and the composition is modified instead, one of the vertical branches of the inversion line is eventually reached and crossed over and the inversion occurs, this time as a consequence of the high internal phase content (horizontal arrows in Figure 1b). These two kinds of morphology swap are called dynamic inversions since they happen as a consequence of a change. Experimental evidence indicates that when the formulation changes by crossing the horizontal branch separating the two normal A regions, the dynamic inversion takes place exactly at the cross over of the standard inversion line, whatever the position and direction of change. Contrariwise, when the vertical branches are crossed due to a continuous change in the water-to-oil ratio, emulsion inversion does not take place im-
4810
Ind. Eng. Chem. Res., Vol. 40, No. 22, 2001
mediately. Instead, it is delayed in both directions beyond the crossing of the standard inversion line. This feature generates wedge-shaped regions called “hysteresis” zones, where any of the two emulsion morphologies may be found depending on the direction of the change (shaded zones in Figure 1b). This “memory” effect can be advantageously used to attain emulsions with specific properties that can be found on either side of the inversion boundaries. Recent studies6,11,19,20 have reported that the extent of these hysteresis zones depends on the emulsification protocol, as well as some characteristics of the system. The formulation-composition mapping of emulsion properties is an organized piece of know-how, which is the framework of the formulation engineering approach of emulsion making.21,22 Programming a path on one side of the inversion line can (1) change all the emulsion properties according to the prevalent features in every location if the changes occur slowly or (2) change only some of them, while keeping others unaffected thanks to the “memory” feature, if a sudden “quench” action is taken, whatever variable is used to produce the rapid modification of formulation or composition. Analysis of the Stirring Pattern Produced by a Rotating Disk At the onset of emulsification, the stirring device starts mixing the liquid in the region located around it. Depending upon the position of the stirrer in the vessel, in particular its position with respect to the oil-water interface, the mixed fluids can contain more water or more oil, thus setting the genesis of emulsification at high or low water-to-oil ratio (WOR), an alternative that could end up in one emulsion type or the other. To discuss the influence of WOR alone, let us start this analysis with a system that only contains water and a low viscosity oil in similar amounts, with no surfactant. The used stirrer device is a rotating disk, which induces only tangential and centrifugal motion to the fluid in which it is located. In a cylindrical vessel, such as a high profile beaker about twice the diameter of the disk, this motion in an homogeneous medium results in a double convection pattern above and below the disk plane, with practically no mixing between the two regions. When the vessel contains two fluids originally separated by a horizontal interface, the effect of the disk rotation depends on its position. In Figure 2 the impeller is located above the interface, i.e., in the oil phase, and it promotes a rolling toroidal pattern between the disk and the interface. The secondary axial motion of the oil phase returning toward the center of the disk results in a deformation of the interface as indicated. If the proper hydrodynamic conditions are met, a vortex is formed and the water phase is dragged into the oil phase against the gravity pull. The sucked water starts being emulsified in the sheared region around the rotating disk as a W/O emulsion containing a small amount of water dispersed in oil. As more and more water is dragged into this vortex, the W/O emulsion occupies a larger volume and limit between the emulsion and the free water moves down. The density difference between the W/O emulsion and the nonemulsified water diminishes, so that the equilibrating gravity pull vanishes and the vortex becomes easier to form. In the sheared region the water drops are incorporated little by little to the already formed W/O emulsion. At some moment, all the water is incorporated and the system becomes macroscopically homogeneous.
Figure 2. Flow pattern and emulsification induced by a rotating disk in a cylindrical vessel (position of the disk above the oilwater interface).
A similar pattern arises if the disk is placed below the interface, but this time an O/W emulsion is produced instead. Thus, it can be said that the continuous phase of the emulsion is the one in which the stirrer is located, since this favors the initial emulsification with a high content of this phase, independently of the other variables that could interfere, such as the formulation variables. This is consistent with previously reported trends.23-25 This phenomenology is rather insensitive to the exact location of the disk inside a phase, because the described phenomenon is an instability of the cusp bifurcation type,1 which keeps growing from the original situation. This is a typical case in catastrophe theory behavior,3 in which a very tiny deviation at the start could end up in considerable changes later on. The relative viscosities of the two fluids has a minor effect, as it will be discussed later. Let’s now consider the case of a complete ternary system containing water, oil and a surfactant, in which the generalized formulation can exhibit hydrophilic, lipophilic or neutral tendencies depending on the case. The stirring experiment previously described applies as well in the presence of a surfactant, but this time the emulsion type could match or not the one that is expected from the formulation point of view. If the final emulsion belongs to the normal A+B+ and A-C- regions, there is no dilemma since both formulation and composition produce the same predisposition. The problem arises when there is a conflict between the two effects, i.e., when the location of the stirrer and the local WOR condition around it, favors the abnormal emulsion type (C+ or B- zones in the bidimensional map). Experimental Results Basic System. The first tested system contains sodium dodecyl sulfate (0.5 wt %), a kerosene cut oil equivalent to heptane (EACN ) 7), and an aqueous solution of NaCl. A certain amount (4.76 vol %) of n-pentanol is added to the system to curtail the extreme hydrophilicity of the surfactant. The phase behavior and
Ind. Eng. Chem. Res., Vol. 40, No. 22, 2001 4811
Figure 3. Shift in inversion line with the initial position of the stirrer.
emulsion properties of this system have been reported elsewhere.9,26 The formulation is modified by changing the salinity of the aqueous phase. The standard inversion line is constructed as in previous studies. Systems presenting a given formulation and composition are first equilibrated at ambient temperature (25 ( 2 °C) and are then emulsified according to a standard procedure that consist in a highly turbulent turbine mixing (Ultraturrax from IKA) placed at the interface and running at 10 000 rpm during 15 s. The emulsion conductivity is then measured to determine its type, either O/W or W/O. The so-called standard inversion line, which is indicated as a bold line in Figure 3 plots, is drawn on the salinity-water fraction map to separate the regions where W/O and O/W emulsions are encountered. It is worth remarking that this label is actually a misnomer, since inversion does not take place when this line is crossed. It is the boundary between the regions where one morphology or the other is attained when emulsifying a preequilibrated system in a very homogeneous and energetic fashion. In the experimental set up to study the effect of the position of the stirrer, the vessel is a high profile glass beaker (6 cm in diameter, 16 cm high). The stirrer impeller is a metallic disk, 3 cm in diameter, presenting four small rounded protuberances (diameter 3 mm, prominence from disk surface 1 mm). In this first type of experiment the disk is placed above or under the interface (typically at 5 mm from interface, but with similar results from 3 mm to 20 mm), and its rotational speed is fixed at 120 rpm. When the disk rotation is started, the mixing process proceeds slowly and the phase that does not contain the disk is dragged into the other according to the mechanism previously discussed. After some time, which can be 1 to 3 min, the system is completely emulsified, and the conductivity is measured to determine the emulsion type. Figures 3 indicate that when the disk is positioned in one phase or the other, the inversion boundary is shifted (see arrows) with respect to the standard inversion line attained in conditions of homogeneous turbulent emulsification. When the disk is located inside the oil phase (Figure 3a), this phase is favored as the external phase of the emulsion according to the previously discussed mechanism. In Figure 3a map it is seen that the A+B+B- zone, where W/O emulsion occurs, stretches out to the right since both vertical branches of the inversion boundary move right. Both normal and abnormal emulsion zones are affected as well, in a very similar way, an indication that this might be only an apparent phenomenon. As a matter of fact, everything happens as if the systems were containing more oil. It
could mean that the region being stirred at the emulsification onset does contain more oil (says 20-25%) than the whole system. For the W/O normal to O/W abnormal catastrophic inversion through the A+/C+ boundary, this means that the inversion point is shifted from 70 to 90% internal water phase or from 30 to 10% external oil phase, which is quite a difference in practical terms. The same effect, but this time in the opposite direction, is found in Figure 3b which corresponds to the location of the stirrer inside the water phase, thus favoring the O/W morphology. It is worth remarking that an extremely high content of internal phase (>95%) can be attained for an O/W emulsion at low salinity, a corroboration that it is quite feasible to prepare a homemade mayonnaise this way. It can be concluded that the location of the stirrer inside a phase results in an additional amount of this phase in the local composition inventory, and thus induces this phase to be the external phase of the initial emulsion. As emulsification proceeds and more internal phase in dragged in the emulsion, the hysteresis feature of the catastrophic dynamic inversion results in the persistence of the initial morphology. It has been known for quite a long time that the zone near optimum formulation exhibits extremely unstable emulsion behavior, and many interpretations have been advanced. Whatever the explanation for this characteristics, this is probably even truer for the abnormal emulsion than for the normal one. As a consequence, the morphology that does not satisfy the formulation should coalesce first, leaving the other one to prevail, though not for a long time after the stirring is interrupted. Though the inversion line location is not available with accuracy near or in the three phase behavior zone (dashed part of the vertical branches) Figure 3 data sustain this statement because the horizontal part on the inversion line is not affected by the position of the stirrer. Other Effects. The generality of this phenomenology was tested by altering some factors whose effects on the standard inversion line have been known for some time.27 In the second study, the system contains the same surfactant and alcohol, but the oil phase viscosity is modified by mixing kerosene with a lubricating oil base. Because of the increased viscosity of the oil, a higher rotational speed (4000 rpm) is used to maintain the drag effect. Whatever the oil viscosity, the first obvious result when Figure 4 is compared to Figure 3 is that the effect of switching the initial position of the disk (from the oil phase to the water phase) is considerably reduced at 4000 rpm stirring with respect to the former case in which the disk speed was 120 rpm. Since the effect of
4812
Ind. Eng. Chem. Res., Vol. 40, No. 22, 2001
Figure 4. Maps showing the inversion line when the stirrer is located inside the oil (black squares) or water (white circles) for different oil viscosity.
the stirrer position is related to the delay in dynamic inversion, this is consistent with the recently reported fact28 that an increase in stirring energy results in a shrinking of the hysteresis zone. Figure 4 also shows the effect of the oil viscosity on the discussed phenomenology. As the oil mixture contains more lube oil and is more viscous, several changes take place: (a) The oil EACN increases and the salinity for optimum formulation decreases to keep HLD ) 0 as expected from eq 2 which holds for many systems.29 It is worth noting that the alcohol content has been slightly reduced with respect to the previous system. According to eq 2, the corresponding reduction in F(A) partially compensates the increase in EACN, thus avoiding to work in the region with extremely low salinity, which is inconvenient with ionic surfactants. (b) The A+/C+ vertical branch of the inversion line is shifted to the left, whereas the B-/A- vertical branch is essentially unmoved, as expected from previous investigations.27 (c) The difference between the inversion line with the disk initial position inside the oil and water phases is not only much smaller than in the previous case, but it is essentially independent of the oil viscosity, despite the considerable shift produced on the A+/C+ branch. Application of the Local WOR Concept What actually happens could be interpreted in a simple way. On one hand, the standard inversion line is considered to be the limit between one type of emulsion morphology and the other when the WOR is kept constant during emulsification. On the other hand, the case in which the stirrer is positioned inside one phase, is a case in which the WOR changes as emulsification proceeds. Thus, the applicable boundary is the dynamic inversion line, which takes place at a higher internal phase content. In cases in which the formulation and composition of the system are represented by points 1, 2, 3, and 4 in Figure 5 a standard emulsification procedure (highly energetic turbulent stirring) would result in a O/W emulsion (abnormal in cases 1 and 2 and normal in 3 and 4) according to their position on the right of the standard inversion branches (vertical straight lines). If the stirrer is initially located in the oil phase, the initial water content of the emulsion is essentially zero (starting point of the arrows), and it slowly increases up to the final value, which corresponds to the overall water content when the system is homogeneous. In cases 1 and 3 the dynamic inversion boundary (bent branch) is not reached and the emulsion keeps the initial W/O morphology, which is normal in case 1 and abnormal
Figure 5. Evolution of emulsified systems taking into account the local WOR as the composition variable on the first emulsion to be formed. Center: case 1. Bottom: case 4.
in case 3. In cases 2 and 4 the emulsion type changes when the path crosses the dynamic inversion line prevailing in the conditions of emulsification to produce an abnormal (respectively normal) O/W emulsion in 2 (respectively 4). Drawings in Figure 5 indicate what happens to the system as emulsification proceeds along arrows 1 and 4. In both cases, the local WOR concept is used as the composition variable until inversion takes place. After inversion and provided that the inverted emulsion is a simple one, then there is only one WOR, i.e., the global one. The local WOR concept applied to the case of multiple emulsions will be discussed in the next section. The previous argumentation may apply not only in the case of very low efficiency mixer, but also in many practical cases of interest whenever the turbulence does not provide instantly a high degree of homogeneity. Even with high-speed turbine blenders, a low mixing region could happen as a consequence of a short-range stirrer pattern, or a high phase or emulsion viscosity. Local WOR may be thus associated with low micromixing occurrence, while high turbulence would be interpreted with overall WOR.
Ind. Eng. Chem. Res., Vol. 40, No. 22, 2001 4813
Figure 6. Use of the local or partial WOR concept to describe the formulation engineering of a multiple emulsion.
The following section applies the local WOR concept to the dual stirring technique to make a multiple emulsion. Interpretation of the Dual Stirring Technique for Multiple Emulsion Making The dual stirring technique which is described in the following paragraphs, allows the making of multiple w/O/W emulsion. In this method there are two stirring devices, each with its range of action. The first one (a short-range small turbine blender) is placed in the oil phase layer and has the initial interface in its range of influence (and then it can sucks in the water phase as illustrated in Figure 2). As the water phase is dragged in the oil as the internal phase droplets, the representative point of the W/O emulsion follows the path 1f3 in Figure 6, where the formulation and WOR refer to local or partial values for this W/O emulsion. As more water goes into the emulsion, the boundary between the emulsion and the nonincorporated water phase moves down, i.e., away from the limited range of action of the first stirrer. Moreover, the increased viscosity of the W/O emulsion with respect to the oil phase, tends to reduce the first stirrer mixing capacity. Alternatively the first stirrer is stopped and the second one started. In this case, the boundary between the emulsion and remaining free water then enters the range of action of the second stirrer, which is a slow helix stirrer that promotes both radial and axial fluid motion over the whole vessel, and whose sucking action favors the other morphology, i.e., O/W. If the composition is taken as the partial WOR of this O/W emulsion, i.e., if the water droplets inside the oil drops are now counted as oil in the volume inventory, then this second emulsification results in a (w/O)/W emulsion along a path such as 4f7 in Figure 6, as far as the WOR change is concerned. This procedure is essentially equivalent to the two step preparation of multiple emulsion, though it requires much less fluid pumping and mixing.30-33 The stirrers can be activated one after the other or both at the same time to favor one or the other mixing process. The formulation can be programmed as well. For instance, a lipophilic surfactant originally dissolved in the oil phase ensures that the first emulsification process takes places in conditions that stabilizes the W/O emulsion (path 1f3). The formulation condition can be thus suddenly changed to HLD < 0 at point 4 just before or at the very moment of the switch from
one type of mixing to the other, so that the formulation reinforces the mixing effect instead of opposing it. This can be done by adding a second surfactant in the free water phase in conditions in which the diffusional transfer restoration of the physicochemical equilibrium could be essentially delayed forever, for instance ,with a viscosity enhancing polymeric surfactant of the hydrophilic type,34 or with a liquid crystal forming surfactant, as a way to stabilize the external emulsion of the multiple emulsion.35 The second emulsification process thus follows the (4f7) path as the oil phase which is actually the w/O emulsion located in (3) is dragged into the (thickened) water. The actual multiple emulsion, that could exhibit a high water content (for instance, in position 8), is more appropriately described by the couple of points 3 and 7 which render the physicochemical and WOR situations for both internal and external emulsions. This example shows another handy use of the concept of local or partial WOR. As times elapses, the diffusional processes always tend to equilibrate the system and the interfacial compositions. It is thus likely that one of the surfactant species is going to migrate quicker that the other, thus affecting one of the emulsion type more than the other. The most affected is the one whose representative point moves first toward HLD ) 0, where a low stability zone is encountered. In the present case and because a water polymer is not likely to migrate through the oil phase, the quickest diffusional process is probably the adsorption of the lipophilic surfactant into the most external emulsion interface. If so, the emulsion which breaks first will be the most external, i.e., the O/W, unless the polymer thickening properties stabilizes the thin film by a kinetic mechanism.36,37 As a final comment, it can be said that if the points representing the two emulsions (3 and 7) are in the A+ and A- stable zones respectively, then the multiple w/O/W emulsion will be stable. If not, or if one of the representative points shifts away and enters the low stability zone near HLD ) 0, then, the multiple emulsion will break up accordingly. The formulation/composition location of both representative points of the multiple emulsion in the map could be thus modified to custom-made the way the system is expected to evolve as time elapses. This is what has been called formulation engineering practice.21,22 Conclusions Because of the memory feature exhibited by emulsions, what happens at the onset of emulsification can be determinant as far as the final emulsion type is concerned. In some instances, the proper location of the stirrer can favor one type of emulsion or the other. The use of the local WOR concept enables the formulator to design complex emulsification processes. Acknowledgment The authors are grateful to the Venezuelan National Research Council CONICIT and to the University of the Andes Research Council CDCHT for sponsoring the Emulsion Science and Technology Research Program at FIRP (Formulation, Interfaces, Rheology and Processes) Laboratory. Literature Cited (1) Dickinson, E. Interpretation of Emulsion Phase Inversion as a Cusp Catastrophe. J. Colloid Interface Sci. 1981, 84, 284.
4814
Ind. Eng. Chem. Res., Vol. 40, No. 22, 2001
(2) Salager, J. L. Phase Behavior of Amphiphile-Oil-Water Systems Related to the Butterfly Catastrophe. J. Colloid Interface Sci. 1985, 105, 21. (3) Salager, J. L. Phase Transformation and Emulsion Inversion on the Basis of Catastrophe Theory. In Encyclopedia of Emulsion Technology; Becher, P.; Ed.; Marcel Dekker: New York, 1988; Vol. 3, Chapter 2. (4) Dickinson, E. Thermodynamic Aspects of Emulsion Phase Inversion. J. Colloid Interface Sci. 1982, 87, 416. (5) Vaessen, G. E. J.; Stein, H. N. The Applicability of Catastrophe Theory to Emulsion Phase Inversion. J. Colloid Interface Sci. 1995, 176, 378. (6) Salager, J. L.; Marquez, L.; Pen˜a, A.; Rondon, M.; Silva, F.; Tyrode, E. Current Phenomenological Know-How and Modeling of Emulsion Inversion. Ind. Eng. Chem. Res. 2000, 39, 2665. (7) Nagata, S. Mixing: Principles and Applications; Halsted Press: New York, 1975. (8) Petit, M. Sensibilidad de una emulsio´n a cambios meca´nicos y fı´sico-quı´micos (Spanish). ChE Thesis FIRP Technical Report # 9105. Universidad de Los Andes: Me´rida, Venezuela, 1991. (9) Salager, J. L.; Min˜ana-Pe´rez, M.; Perez-Sanchez, M.; RamirezGouveia, M.; Rojas, C. Surfactant-Oil-Water Systems near the Affinity Inversion-Part III: The Two Kinds of emulsion inversion. J. Dispersion Sci. Technol. 1983, 4, 313. (10) Dickinson, E. Emulsions. In Annual Reports C; The Royal Society of Chemistry: London, 1986; p 31. (11) Brooks, B. W.; Richmond, H. N. Dynamics of LiquidLiquid-Phase Inversion Using Nonionic Surfactants. Colloids Surf., A 1991, 58, 131. (12) Davis, H. T. Factors Determining Emulsion Type: HLB and Beyond. Colloids Surf., A 1994, 91, 9. (13) Binks, B. P.; Lumsdon, S. O. Catastrophic Phase Inversion of Water-in-Oil Emulsions Stabilized by Hydrophobic Silica. Langmuir 2000, 16, 2539. (14) Min˜ana, M.; Jarry, P.; Pe´rez-Sanchez, M.; RamirezGouveia, M.; Salager, J. L. Surfactant-Oil-Water Systems near the Affinity Inversion-Part V: Properties of Emulsions. J. Dispersion Sci. Technol. 1986, 7, 331. (15) Salager, J. L.; Ma´rquez, N.; Graciaa, A.; Lachaise, J. Partitioning of Ethoxylated Octylphenol Surfactants in Microemulsion-Oil-Water Systems. Influence of Temperature and Relation between Partitioning Coefficient and Physicochemical Formulation. Langmuir 2000, 16, 5534. (16) Salager, J. L.; Microemulsions. In Handbook of DetergentsPart A: Properties; Broze, G., Ed.; Marcel Dekker: New York, 1999; Chapter 8, p 253. (17) Bancroft, W. D. The Theory of Emulsification. J. Phys. Chem. 1913, 17, 501. (18) Salager, J. L. Emulsion Properties and Related Know-How to Attain Them. In Pharmaceutical Emulsions and Suspensions; Nielloud, F., Marti-Mestres, G., Eds.; Marcel Dekker: New York, 2000; Chapter 3, p 73. (19) Brook, B; Richmond, H. N. Phase Inversion in Non-Ionic Surfactant-Oil-Water Systems, II. Drop Size Studies in Catastrophic Inversion with Turbulent Mixing. Chem. Eng. Sci. 1994, 49, 1065. (20) Silva, F.; Pen˜a, A.; Min˜ana-Pe´rez, M.; Salager, J. L. Dynamic Inversion Hysteresis of Emulsions Containing Anionic Surfactants. Colloids Surf., A 1998, 132, 221.
(21) Salager, J. L.; Ma´rquez, L.; Mira, I.; Pen˜a, A.; Tyrode, E.; Zambrano, N. B. Principles of Emulsion Formulation Engineering. In Surfactants in Solution; Mittal, K. L., Shah, D. O., Eds.; Marcel Dekker: New York. In press. (22) Salager, J. L.; Bricen˜o, M. I.; Bracho, C. L. Heavy Hydrocarbon Emulsions-Making Use of the State of the Art in Formulation Engineering. In Encyclopedic Handbook of Emulsion Technology; Sjo¨blom, J., Ed.; Marcel Dekker: New York, 2001; p 455. (23) Quinn, J. A.; Sigloh, D. B. Phase Inversion in the Mixing of Immiscible Liquids. Can. J. Chem. Eng. 1963, 41, 15. (24) Selker, H. A.; Sleicher, C. A. Factors Affecting Which Phase will Disperse when Immiscible Liquids are Stirred Together. Can. J. Chem. Eng. 1965, 43, 298. (25) Rodger, W. A.; Trice, V. G.; Rushton, J. H. Effect of Fluid Motion on the Interfacial Area of Dispersions. Chem. Eng. Prog. 1956, 52, 515. (26) Salager, J. L.; Loaiza-Maldonado, I.; Min˜ana-Pe´rez. M.; Silva, F. Surfactant-Oil-Water Systems near the Affinity Inversion-Part I: Relationship between Equilibrium Phase Behavior and Emulsion Type and Stability. J. Dispersion Sci. Technol. 1982, 3, 279. (27) Salager, J. L.; Lo´pez-Castellanos, G.; Min˜ana-Pe´rez. M. Surfactant-Oil-Water Systems near the Affinity Inversion-Part VI: Emulsions with Viscous Hydrocarbons. J. Dispersion Sci. Technol. 1990, 11, 397. (28) Pen˜a, A.; Salager, J. L. Effect of Stirring Energy upon the Dynamic Inversion Hysteresis of Emulsions. Colloids Surf., A 2001, 181, 319. (29) Salager, J. L.; Morgan, J.; Schechter, R. S.; Wade, W. H.; Va´squez, E. Optimum Formulation of Surfactant-Oil-Water Systems for Minimum Tension and Phase Behavior. Soc. Petr. Eng. J. 1979, 19, 107. (30) Grossiord, J. L., Sellier, M., Eds. Multiple Emulsions: Structure, Properties and Applications; Editions La Sante´: Paris, 1998. (31) Matsumoto, S.; Kita, Y.; Yonezawa, D. An Attempt at Preparing Water-in-Oil-in-Water Multiple Phase Emulsions. J. Colloid Interface Sci. 1976, 57, 353. (32) Matsumoto, S.; Kang, W. W. Formation and Applications of Multiple Emulsions. J. Dispersion Sci. Technol. 1989, 10, 455. (33) Matsumoto, S.; Koh, Y.; Michiura, A. Preparation of w/o/w Emulsions in an Edible Form on the Basis of Phase Inversion Technique. J. Dispersion Sci. Technol. 1985, 6, 507. (34) Py, C.; Rouvie`re, J.; Tadros, Th. F.; Taelman, M. C.; Loll, P. Investigation of Multiple Emulsion Stability Using Rheological Measurements. Colloids Surf., A 1994, 91, 215. (35) Kavaliunas, D. R.; Frank, S. G. Liquid Crystal Stabilization of Multiple Emulsions. J. Colloid Interface Sci. 1978, 66, 586. (36) Ivanov, I. B.; Kralchevsky, P. A. Stability of Emulsions under Equilibrium and Dynamic Conditions. Colloids Surf., A 1997, 128, 155. (37) Ivanov, I. B., Ed. Thin Liquid Films; Marcel Dekker: New York, 1988.
Received for review February 26, 2001 Revised manuscript received August 22, 2001 Accepted August 23, 2001 IE010196R