Miniemulsions: overview of research and applications.
Keywords: Miniemulsion, Ostwald ripening, enhanced nucelation, hybrid latexes, encapsulation, droplet and particle size distributions, reaction kinetics, reaction mechanisms
The deliberate preparation and polymerization of what later became known as miniemulsions was undertaken at Lehigh University in 1972 by a group consisting of Professor John Vanderhoff, Visiting Professor John Ugelstad, and Mohamed El-Aasser who had just arrived at Lehigh as a post doc. It was noted in a lab book entry made by Vanderhoff (dated, signed, and witnessed June 19, 1972; see Figure 1) where he wrote: "These ideas come from the earlier observation of Dr. Ugelstad that, with certain combinations of anionic emulsifier and fatty alcohol (e.g., sodium lauryl sulfate and cetyl alcohol) used in emulsion polymerization vinyl chloride, the monomer emulsion droplets are a significant loci for initiation of polymerization." In fact, by the time this record was made, laboratory experiments at Lehigh had already been carried out by Mohamed El-Aasser showing that "styrene monomer emulsion droplets smaller than 0.2 [micro] could be prepared in this way." He polymerized these emulsions with results indicating that these droplets too were "a significant loci for polymerization initiation." Proposed uses of this new technology included the preparation of artificial latexes with small particle size and also the encapsulation of pigment particles by polymerization of monomer in the form of small droplets containing the pigment.
So it seems that a curious observation led to initiation of fundamental and applied research in miniemulsions. Note that high shear was not mentioned here, as is the standard for preparing miniemulsions today. In fact, in much of the early work employing cetyl alcohol (CA) as the cosurfactant, miniemulsions were prepared by simple mixing of the monomer into the surfactant/cosurfactant aqueous "gel" phase. (1-3) This led to much research into the origin of the ability of this combination to create such small droplets. (4-6) As high shear homogenization became standard (and was required in many cases), the early history became all but forgotten. As another historical note, the term "cosurfactant" originated in this early work using cetyl alcohol, a true cosurfactant. Later, when other materials such as hexadecane were used, the term cosurfactant was held over even though it has no surface active properties. Today, the term "costabilizer" has been adopted as a more accurate descriptor of the role of this agent in the miniemulsion.
It is not the intent of this article to provide a comprehensive review of miniemulsions over the past 30 years. This has largely been accomplished in several major reviews appearing in the last two years. (7-9) In these, the authors have noted the recent acceleration in the number of publications highlighting some aspect of miniemulsions, ranging from the fundamental to the applied, particularly in the five years since the last major reviews appeared (see Figure 2). (10,11) It should be noted that the emphasis in these reviews has been on the polymerization of miniemulsions, while the issue of emulsification of polymer solutions to create artificial latexes, an important subset of miniemulsions, is more rarely covered, (11) as much less research has been exclusively carried out in this area. Still, one of the earliest commercialized products making use of miniemulsification technology was Aquacoat[R] (FMC), the latex form of cellulosic materials used today to coat pharmaceutical pills.
[FIGURE 1 OMITTED]
Although a number of academic and industrial research groups have engaged in miniemulsion research, the Emulsion Polymers Institute of Lehigh University has persistently led the field in terms of publications, particularly with respect to the fundamental understanding of the nature of miniemulsions, their preparation, polymerization, and application. Since the first Ph.D. dissertation was written on the subject, a total of 22 Ph.D. and 16 M.S. degrees have been completed on some aspect of miniemulsions. In addition, seven visiting scientists and professors have engaged in this research while at Lehigh, as well as over 20 other collaborators from various academic and industrial research labs around the world. This research has favored the preparation of monomer miniemulsions and their polymerization to produce synthetic latexes with roughly 65% of the work being assigned to this while polymer emulsification studies producing artificial latexes constitute about 20% of the work, with the remaining roughly 15% devoted to understanding emulsification in the CA costabilized systems.
This article will first focus on some of the basic issues regarding miniemulsions with respect to their preparation and polymerization, then note some of the advancements, and finally look at some applications.
POLYMERIZATION OF MINIEMULSIONS
The most vexing question clouding miniemulsion polymerization research has been: how many of the monomer droplets present in the initial miniemulsion become polymer particles? Similarly, how many particles result from radical polymerization in the droplets? The first question implies that in many cases not all droplets become particles (i.e., the number of initial droplets is greater than the number of final particles ([N.sub.d.sup.o]>[N.sub.p.sup.f])). The second question implies other sources of particle formation, typically homogeneous nucleation, which can result in both [N.sub.d.sup.o]>[N.sub.p.sup.f] or [N.sub.d.sup.o]<[N.sub.p.sup.f]. A number of methods have been applied to determine the reality with varying degrees of success. However, the question largely remains in most cases. Why? Because we are unable to trace the number of droplets as they become particles. We do not have an accurate and reliable means of monitoring the droplet size distribution (DSD) from the moment of initiation to its final disappearance sometime during the polymerization. This is the Achilles' heal of miniemulsion polymerization research. An equally important question regarding the droplets is how their distribution changes during and after the emulsification process. This question can only be understood by reviewing some of the background regarding miniemulsions, their preparation and stabilization.
Preparation of Miniemulsions
Today, most oil-in-water miniemulsions are generally prepared in a similar manner by subjecting the system to high shear. The system is usually comprised of an oil phase containing the monomer(s) and costabilizer, and sometimes a polymer, initiator, colloid (as in encapsulation), and others. The water phase contains the stabilizer, which is often an anionic surfactant such as SLS, but can be a cationic, nonionic, or polymeric species as well. The role of the costabilizer, as the name suggests, is to act in concert with the stabilizer to provide stability to the droplets, in this case, stability against extensive diffusional degradation (Ostwald ripening). Shear is most often applied after a crude emulsion is created by simple mixing of the ingredients as with a magnetic bar or impeller. Sonifiers, Microfluidizers, and Manton-Gaulin homogenizers are the most common devices for applying the high shear necessary to break up the oil phase, although a number of other devices have also been used. Sometimes a combination of these is applied (e.g., one minute sonification followed by 10 passes through the Microfluidizer). (12) What happens during and after this process is complex as many processes are set in motion simultaneously. In using these high energy devices, the large monomer droplets (many microns in size) of the crude emulsion are broken into smaller sizes by mechanisms that have been described in terms of cavitation, shear, and impact. Some energy is dissipated as heat and thus, often the instrument and/or accompanying storage vessels are chilled with ice baths to prevent excessive heating and increased degradation of the miniemulsion.
With the creation of smaller droplets, surfactant adsorbs at the oil/water interface as it seeks its own equilibrium adsorption. In some cases, the costabilizer (such as CA) is also present at the interface and interacts with the surfactant. This interaction has been investigated in a number of studies. (2-6) Other ingredients also seek to minimize their free energies. The monomer diffuses from smaller to larger droplets (Ostwald ripening). If not for the costabilizer, the system would return to a state of relatively large oil droplets soon after the shear is removed. However, the low water solubility of the costabilizer causes it to concentrate in the smaller droplets and become diluted in the larger ones. The redistribution of monomer is governed by thermodynamics and the effectiveness of the costabilizer is determined by its water solubility and molecular weight. Its role cannot be overemphasized. More on this will be discussed later.
A distribution of droplet sizes having a distribution of compositions now exists. With further shear, larger droplets are further broken down while some small droplets are lost by collisions (and coalescence) with other droplets. The dynamics of droplet breakage and coalescence seems to approach a steady-state after a certain amount of time or passes have been sustained. At this stage, further shear is not likely to decrease the average droplet size or change the distribution, which is expected to be relatively broad. The high pressure homogenizers are considered to result in narrower DSDs than the sonifier based on the uniformity of the exposure of all elements of the emulsion to the same shear conditions.
[FIGURE 2 OMITTED]
When the emulsification is terminated, the miniemulsion will not be in a static state, but will be subject to further redistribution of components seeking equilibrium and lowered free energies. Diffusion of monomer, coalescence of droplets, and redistribution of surfactant is likely but the extent and importance of each are not typically known and would vary from system to system. A pseudo-equilibrium is expected where, in a time frame of hours or days, little changes in the miniemulsion, allowing it to be subjected to free radical polymerization, thereby creating a synthetic latex. It is a metastable state, as illustrated in Figure 3, lying between conventional emulsions regarded as unstable and microemulsions, which are thermodynamically stable. (10)
[FIGURE 3 OMITTED]
As stated above, the role of the costabilizer is critical in determining the stability against diffusive degradation of a miniemulsion. The surfactant provides the crucial colloidal stability component. The following discussion will be limited to costabilizers, such as hexadecane, which are not present at the interface and do not exhibit substantial interactions with the surfactant at the interface (as cetyl alcohol).
The basis of miniemulsion stabilization against diffusional instability is thermodynamics as described by the well-known expression for the partial molar Gibbs free energy of monomer contained in a droplet comprised of the monomer and a small amount of the water-insoluble costabilizer (13,14):
[bar.[DELTA][G.sub.m]]/[RT] = ln([[phi].sub.m]) + (1 - [m.sub.mc])[[phi].sub.c] + [X.sub.mc]([[phi].sub.c])[.sup.2] + [[2[bar.V.sub.m[gamma]]]/[rRT]] (1)
where [[phi].sub.m] and [[phi].sub.c] are the volume fractions of the monomer and costabilizer, [m.sub.mc] is the ratio of molar volumes of the monomer and costabilizer (= [bar.V.sub.m]/[bar.V.sub.c] = 1/[j.sub.2]), [[chi].sub.mc] is the Flory interaction parameter between the monomer and the costabilizer, [gamma] is the droplet/water interfacial tension, [gamma] is the radius of the droplet, R is the gas constant, and T is the temperature. At equilibrium, this equates to zero. The stabilizer is required to have two characteristics to be effective: low water solubility and low molecular weight. Use of the above equation in fact assumes no water solubility of the costabilizer. In practice, its water solubility must be sufficiently low that significant redistribution of the costabilizer by diffusion between droplets does not take place in the time scale of the experiments. The second requirement of low molecular weight will be illustrated by using three simplified examples.
[FIGURE 4 OMITTED]
SWELLABILITY OF DROPLETS OF COSTABILIZER: The power of the costabilizer to retain, or in this case, to take in monomer (styrene in this example), can be illustrated by considering the swelling capacity of a given size droplet (100 nm) comprised solely of the costabilizer and having a constant interfacial tension (5 dyn/cm). If this droplet, contained in an aqueous solution of the surfactant, has access to an infinite reservoir (pool) of styrene monomer, it will swell until equilibrium (pseudo) is reached. Table 1 reports the swelling capacity of such droplets as function of the value of [j.sub.2] of the component being swollen (the droplet or particle, depending on the material), the potential miniemulsion costabilizer (T = 27[degrees]C; [[chi].sub.mc] = 0.5). (14) It is clear that the swelling capacity ([V.sub.f]/[V.sub.i]) decreases sharply with an increasing [j.sub.2]. Note that hexadecane, considered an optimum costabilizer, has a [j.sub.2] value of 2.64. Lower values would result in higher water solubilities, negating the long-term effectiveness of the costabilizer. A high molecular weight polymer would have a [j.sub.2] approaching infinity and result in a low swelling capacity. This is further illustrated in the next example.
SWELLABILITY OF DROPLETS CONTAINING COSTABILIZER: Now consider a situation in which a "miniemulsion" droplet of a specific size and containing 4% by volume of a "costabilizer" (with the remainder being styrene monomer) is dispersed in an aqueous phase where the temperature and interfacial tension are constant at 25[degrees]C and 3.4 dyn/cm, respectively. This is similar to the situation after the preparation of a typical miniemulsion. Now again contact the emulsion (containing all the same size and composition droplets) with an infinite reservoir or pool of styrene monomer and determine the equilibrium conditions attained by varying the initial droplet size (50 and 100 nm) and the chain length of the "costabilizer," in this case, oligomers of polystyrene. The results are illustrated in Figure 4 as the equilibrium diameter of the droplets as a function of the chain length or [j.sub.2] of the oligomers. For the 100 nm initial droplet size, the droplets will take in additional monomer if [j.sub.2] is less than 6 (MW < 624); otherwise for longer molecules, monomer will diffuse to the reservoir. For the 50 nm droplets, [j.sub.2] should be even smaller ([less than or equal to] 3) to prevent monomer loss. Note that in both cases, hexadecane (diamonds) allows for an increased uptake. This shows that the requirements for a good costabilizer are narrow and that polymer of any significant degree of polymerization itself cannot act thermodynamically as a miniemulsion costabilizer. Of course, this situation is rather simplified as compared to the reality of broad droplet size distributions in miniemulsions. The next example will take this one step closer to this situation.
DROPLETS IN EQUILIBRIUM: If droplets of different size (20, 40, and 60 nm diameters) but the same composition are allowed to equilibrate with each other, what will be their extents of growth or shrinkage from redistribution of the monomer (styrene)? Figure 5 shows the variation in the resulting droplet volumes (represented as the ratio of the equilibrium volume to the initial volume) for the three droplets as a function of the volume fraction of the "costabilizer" initially present in all droplets ([[phi].sub.20]). (15) Hexadecane and polystyrene (MW = 100,000 g/mol) are the two costabilizers. The interfacial tension was 5 dyn/cm and considered constant at all sizes and compositions (a simplification). At an initial 0.04 volume fraction (4%) of HD as the costabilizer, the two smaller droplets shrink some while the larger one grows at their expense. This is an Ostwald ripening effect, as described earlier, but is substantially limited by the hexadecane. The droplets containing polystyrene shrink to a much greater degree, again illustrating that this material cannot play the role of a true costabilizer.
The extension of these examples to real miniemulsions is obviously more complex but doable. However, without better information on the droplet size and composition distributions, this would be an unproven exercise.
Miniemulsions are usually not an end in themselves but rather a means to an end. Artificial and synthetic latexes are the result of the next process step, which is either the removal of the solvent from miniemulsions prepared from polymer solutions or polymerization of monomer miniemulsions, respectively. The focus here is on the latter.
Much work has been published reporting the polymerization of miniemulsions. As mentioned earlier, the weakness in this area has been in the understanding of the transition from a droplet size distribution (DSD) to a particle size distribution (PSD). Usually the final particle size is reported as averages or a distribution obtained by the normal means of characterizing latexes. The droplet size may also be reported as an average obtained most frequently by dynamic light scattering ([D.sub.n], [D.sub.v]; number and volume-average diameters) or more recently by soap titration ([D.sub.vs]; volume-surface average diameter). (16) However, little is really known about this transition. Interestingly, a wide range of PSDs have been reported for the polymerization of miniemulsions. Figure 6 illustrates the main possibilities. Narrow PSDs have been reported infrequently. Early in the development of miniemulsion latexes, Hansen and Ugelstad found conditions resulting in polystyrene (PS) latexes with distributions having coefficients of variation (standard deviation ([sigma])/[D.sub.n]) of 5% and above. (17) The former is considered narrow but not monodisperse by the strictest definition (i.e., [sigma]/[D.sub.n] [less than or equal to] 2%).
More commonly, coefficients of variation of 10 to 20% have been reported, (18,19) which is also not uncommon for latexes made by conventional emulsion polymerization. Most recently, we have seen narrow PSDs produced in miniemulsion homopolymerizations of styrene and n-butyl methacrylate (BMA) when using the hydrogen peroxide ([H.sub.2][O.sub.2])/ascorbic acid (H2A) redox pair. (15,20) Here it is reasoned that the burst of radicals produced by this pair nucleates all the droplets in a short time (low conversion), thus allowing for a period of particle growth that is long relative to the nucleation period. This is one criterion often mentioned for achieving latexes of narrow distribution. This is supported by the constant number of particles with increasing initiator concentration: all droplets and only droplets lead to particles. (20)
The miniemulsion "ideal" is not only that all droplets become particles but that there is what some authors refer to as a 1-to-1 copying of droplets to particles. This situation is rare if it exists at all. For this to happen, no transfer of monomer between droplets/particles should occur throughout the entire polymerization. Monomers with any water solubility will diffuse in response to imbalances in the thermodynamic equilibrium between particles. This imbalance occurs readily because of the nature of emulsion polymerization. In the size range of particles being produced, "zero-one" systems are common. This refers to the number of radicals (n) growing in an individual swollen particle at any given time. Because some particles contain a growing radical (n = 1) and some do not (n = 0), this creates a thermodynamic imbalance of monomer as it is consumed in the former particles. This monomer must diffuse from those where no radical polymerization is occurring (the latter) to those where it is occurring. Of course, particles switch between the zero and one states often during a polymerization as can be roughly estimated by the number of chains in any given particle at the end of the reaction. Dead (non-growing) chains are created either when termination of two radicals occurs (combination or disproportionation) or when chain transfer occurs from the growing chain to another molecule such as the monomer itself. Therefore, it is clear that a 1-to-1 copying cannot occur with this continuous transport of monomer among the growing particles.
[FIGURE 5 OMITTED]
There may be a rare case where the 1-to-1 scenario might be approached and that is for monomers with exceedingly low water solubilities. One such example is octadecylmethacylate (ODMA). Its water solubility has been shown to be sufficiently low that it can serve as an effective costabilizer. (21) It can also be both the primary monomer and "costabilizer." Some miniemulsion polymerization results obtained in our labs for this monomer are reproduced in Figure 7 where the evolution of the polymerization rate as a function of conversion is shown for three different "miniemulsions." (22) Two "classical" recipes using CA/SLS and HD/SLS are compared to one with no additional costabilizer (i.e., ODMA/SLS alone). Little difference is seen in the rate of polymerization curves. They rise to a rate maximum at about 15% conversion and then steadily decrease. Three sets of droplet/particle size data ([D.sub.n], number-average diameter) are presented but are not distinguishable. The size was obtained by capillary hydrodynamic fractionation (CHDF, Model 1100, Matec Applied Sciences), which normally has much difficulty in determining the droplet size distributions of miniemulsions. (23) In this case, the low water solubility of the monomer allowed for the measurement and the particle size represents the swollen particles rather than the "dry" particles as is typically obtained when analyzing particle size distributions of samples taken during a reaction. Within the scatter, the size is constant from the all-droplet beginning to the all-particle end, implying that all droplets become particles basically maintaining their original identity. This sort of data is rarely reported.
[FIGURE 6 OMITTED]
The third case illustrated in Figure 6 is that resulting in a broad particle size distribution, even broader than the original droplet size distribution. Although this has long been the expectation of most miniemulsion polymerizations, hard evidence is scant. As an extreme case, the stable free radical polymerization (SFRP) of a styrene miniemulsion at 125[degrees]C resulted in a latex with a polydispersity index (PDI) of 1.75, considered to be an extremely broad distribution. (24) More typically, long nucleation periods, as seen in styrene/CA/SLS systems, (18,25) where only a fraction of the droplets is nucleated, are considered partly responsible for this broadening resulting in PSDs having PDIs of about 1.03-1.06, where the coefficient of variation is in the range of 12 to 20%. Other mechanisms of broadening include limited aggregation, creating larger particles, and homogeneous nucleation, creating smaller particles, which are considered likely in the SFRP case mentioned above. The extent of each of these is not easily determined without knowledge of the droplet size distribution and the evolution of the resulting PSD. In the S/CA/SLS system, an increasing number of particles with increasing initiator is cited as an indication of the relatively low fraction of droplets nucleated. (18) Such results are shown by the lower curve in Figure 8 where the number of particles ([N.sub.p]) increased in proportion to the 0.31 power of the KPS concentration. (25,26) The upper curve shows a more interesting case where [N.sub.p] was unaffected by the initiator concentration. (26) Here a small amount of polymer was preadded to the styrene monomer phase prior to miniemulsion formation. This is the subject of the next section.
'Enhanced' Droplet Nucleation
As part of the study aimed at understanding what was considered to be the low "nucleation efficiency" of droplets compared to particles (radical entry efficiency) in seeded emulsion polymerizations, polymer was added to the monomer phase (SLS/CA as stabilizer/costabilizer) prior to forming the miniemulsion so as to create "polymer particles" at low conversion (e.g., 1%). The idea was to try to roughly estimate the radical entry rate in this system to see if the polymer increased the ability of the droplet/particle to retain the entering free radical without subsequent desorption. What was found, however, was a vastly increased rate of polymerization and number of particles compared to the miniemulsion polymerization with no added polymer. This was unexpected and became known as "enhanced" droplet nucleation. Some of these results are reproduced in Figure 9, which shows an increasing polymerization rate with increasing amount of polystyrene polymer added to the styrene monomer. (27) With increasing initiator (KPS) concentrations at a fixed amount of added polymer (1 wt% on monomer), the number of particles was found to be high and constant (see Figure 8). The particle size distributions revealed a bimodal PSD at low initiator levels, which became unimodal at higher levels. (26) This was taken to mean that although all "droplets" (containing polymer) became particles, some fraction, inversely dependent on the initiator concentration, remained unentered by a free radical during the subsequent polymerization until sufficient free radicals were produced to effect entry of all droplets.
What was responsible for this apparent "enhanced" droplet nucleation? Several theories were proposed (28) and subsequently subjected to testing. (29-31) One has survived. The first idea, which goes back to the original one of adding the polymer, was that the presence of polymer brought about a more efficient capture of free radicals by increasing the internal viscosity of the droplets (i.e., reduced diffusivity), allowing time for propagation of the oligomer chain, resulting in irreversible capture. The test was to add polystyrene polymer of varying molecular weight and viscosity and see how the polymerization kinetics were affected. In fact, no discernable effect was found. All reactions produced the same rate of polymerization profiles and the same number of particles (within experimental error). (29) A viscosity effect was not seen.
The second idea was that the presence of the polymer in the droplets acted to disrupt the ordered structure of the SLS/CA at the droplet/water interface, allowing radicals to more freely enter the droplets. The polymer used in the prior studies had been obtained with sulfate end groups (as produced in a conventional emulsion polymerization). The thought was that these hydrophilic end groups would prefer to protrude through the interface into the surrounding aqueous environment and thereby create flaws in the ordered SLS/CA structure. (4-6) Therefore, several types of polystyrene were prepared having different chain ends (H, S[O.sub.3], S[O.sub.4]) and molecular weights (39,000-103,000 g/mol), and these were used to prepare styrene miniemulsions (1 wt% PS). Once again, all reactions resulted in the same polymerization rate profiles and number of particles (within experimental error). (29) Surface disruption did not seem to be the issue.
The third proposed explanation for the "enhanced" nucleation was that the presence of the polymer acted to ensure the existence of all droplets, disallowing any disappearance by monomer diffusion. Since the polymer was too large to have any water phase solubility, it would have to remain in its original droplet unless it "disappeared" by a colloidal instability mechanism. This theory has survived most of the subsequent testing including variations in the composition of the polymer (poly(n-butyl acrylate), poly(2-ethylhexyl acrylate), poly(lauryl acrylate), poly(n-octyl acrylate), and poly(vinyl acetate); all of about 100,000 g/mol molecular weight) in styrene miniemulsions and variations in the composition of the monomer phase (n-butyl acrylate, styrene/n-butyl acrylate, and methyl methacrylate with various added polymers). (15) The "enhanced" nucleation effect was equal within experimental error in all comparable systems.
Results were also found in some of these studies, however, which seemed to be inconsistent with the idea of droplet preservation by the added polymer. Figure 10 shows such results where the molecular weight and amount of added polystyrene were varied in styrene/CA/SLS miniemulsions. (15) It was expected that, provided some polymer was present in each droplet, each would be preserved. So it would not matter if, in one miniemulsion, each droplet contained 300 chains on average of a 1,000 g/mol PS while in another, each droplet contained 300 chains of a 100,000 g/mol PS. The only variation thought likely here was that the higher molecular weight and thus higher polymer content for a fixed number of chains might cause fewer droplets to be produced owing to the higher viscosity of the S/PS solution. This then might lead to slower polymerization kinetics resulting from fewer particles. This was not found to be the case, as seen in Figure 10. In fact, the 1% 100,000 g/mol PS produced a much higher reaction rate compared to the 0.01% 1,000 g/mol PS even though similar droplet sizes should have contained similar numbers of chains, all insoluble in the aqueous phase. The particle numbers were consistent with the polymerization rates. The question remains as to why the lower molecular weight polymer resulted in fewer particles being formed and is this a reflection on the amount of droplets preserved. These results are nonetheless intriguing, with more work being required to reach a decisive conclusion.
Other Applications of Added Polymer
Adding polymer to the miniemulsion oil phase has been utilized in a diverse range of studies. Three described here include: (1) controlled radical polymerization; (2) preparation of hybrid latexes; and (3) encapsulation of inorganic particles. A summary of these findings are presented.
CONTROLLED RADICAL POLYMERIZATION: Research in the preparation of relatively narrow molecular weight polymers by several controlled radical polymerization mechanisms has been intensive for nearly 10 years now. One of these is referred to as stable free radical polymerization (SFRP), which, as the name implies, employs a stable free radical that cannot initiate polymerization by itself but engages in a reversible combination with a propagating polymer chain. TEMPO (2,2,6,6-tetramethyl piperidenyl-1-oxy) is the most frequently used SFRP agent as it is commercially available. Initially, bulk polymerizations (primarily styrene) were studied, but interest in preparing colloidal particles by emulsion polymerization eventually followed. Unstable latexes were generally produced, attributed to the high polymerization temperature required for TEMPO to be effective in regulating the molecular weight (115[degrees]C-140[degrees]C). Miniemulsion polymerization was soon adopted as a means of preparing stable latexes (~20% solids) containing polymers having relatively narrow molecular weight distributions (MWD). (32)
[FIGURE 7 OMITTED]
To gain a more thorough mechanistic understanding of these nitroxide-mediated living free radical miniemulsion polymerizations, TEMPO was replaced by an initiating species comprised of short polystyrene chains capped by TEMPO, these being referred to as TTOPS (TEMPO-terminated oligomers of polystyrene). (24,33,34) The main purpose here was to fix a known number of initial chains (used as a macroinitiator), which would subsequently grow in miniemulsion polymerizations. The basic process occurring in the droplet/particles is illustrated in Figure 11, where the macroinitiator in this illustration has a molecular weight of about 1500 g/mol (TTOPS 1500) or a degree of polymerization of about 13 (the newly added mers can be styrene of another monomer such as n-butyl acrylate). Unlike ionic polymerizations, termination still plays a significant role in these polymerizations producing some "dead" polymer and broadening the MWD (PDI TTOPS [approximately equal to] 1.14-1.24; final polymer PDI [approximately equal to] 1.3-1.8, depending on conditions). Although these molecular weight distributions are quite narrow relative to what is normally produced by conventional free radical miniemulsion polymerization, the particle size distributions are much broader as mentioned earlier (PSD PDI [approximately equal to] 1.8) with sizes ranging from 20 nm up to 500 nm in a single latex. The cause of this wide distribution is likely a combination of some colloidal instability at the high reaction temperature (125[degrees]C) creating the largest particles and some additional particle formation by homogeneous nucleation (thermal initiation and water solubility of TEMPO allow creation of new particles with controlled molecular weight).
[FIGURE 8 OMITTED]
In these SFRP reactions, the rate of polymerization is slow and largely independent of the number of particles in the system. (34) The low number of active free radicals in the system is considered responsible where under typical conditions only one in 300 particles has an active growing radical at any given time (n[approximately equal to]0.003). In normal miniemulsion polymerizations, the average number of radicals per particle would be close to one-half where one in every two particles would contain a growing polymer radical.
The miniemulsion method using TTOPS has also been extended to the preparation of some block copolymers with n-butyl acrylate as the second monomer. (35) Interestingly, a population of dormant TTOPS exists, which decreases with increasing amount of ascorbic acid added as a radical scavenger. It is not clear why this is the case, requiring further studies.
HYBRID LATEXES: By polymerizing a monomer in the presence of a polymer having a differing chemical composition, hybrid materials can be made with properties unique to the system. This can readily be accomplished by applying the miniemulsion technique resulting in hybrid latex particles.
[FIGURE 9 OMITTED]
Polyurethane/Acrylic--Recently, small-sized (~50 nm) polyurethane/poly(n-butyl methacrylate) [(PU)/(PBMA): 25/75] hybrid latex particles were prepared by miniemulsion polymerization. (20) The [H.sub.2][O.sub.2]/ascorbic acid redox pair was used to effect polymerization at 30[degrees]C, where Ostwald ripening of the miniemulsion could be minimized. Urethane prepolymers (both isophorone diisocyanate (IPDI) and methylene-di-p-phenyldiisocyanate (MDI) based) were prepared containing a small amount of grafting agent (hydroxy ethyl methacrylate, HEMA) to promote compatibility of the PU and PBMA. These prepolymers were mixed with the BMA and HD (costabilizer) prior to sonification to form the miniemulsion. Polymerizations were fast, high conversions being achieved within 15 min, considerably faster than control reactions in the absence of the urethane prepolymer. Although this might sound like an "enhanced" nucleation effect caused by the presence of the urethane prepolymer, in reality, finer miniemulsion droplets ([D.sub.w] [approximately equal to] 50 nm, PDI [approximately equal to] 1.17; dynamic light scattering) were produced with the urethane/BMA/HD mixture compared to the simpler BMA/HD ([D.sub.w] [approximately equal to] 82 nm, PDI [approximately equal to] 1.07 nm). This was attributed to the hydrophilicity of the prepolymer, allowing a more efficient use of the stabilizer (SLS) producing more and thus, smaller droplets.
The urethane prepolymer was chain extended successfully using an oil-soluble chain extender (bisphenol A) and the properties of the resulting stable latexes subsequently determined. The polymer recovered from the latexes was subjected to dynamic mechanical analysis and stress-strain tests. These indicated mixing of the PU and PBMA at the molecular level, while control samples comprised of PU/PBMA blends showed the expected two-phase behavior. The presence of noncrosslinked PU increased the maximum strain of the films significantly while only sacrificing some strength, which was not the case for the blends where both the strength and maximum strain were decreased.
Kraton/Polystyrene Hybrid Latexes--Rubber toughened materials have long been made by creating rubbery inclusions in brittle polymers such as polystyrene (e.g., high impact polystyrene, HIPS). The size of these inclusions can be relatively large and are largely controlled by the mixing conditions during their preparation. As an alternative methodology, miniemulsification of a rubber (Kraton[R] D1102, a styrene-butadiene-styrene triblock copolymer) and styrene monomer mixture in water stabilized by a combination of SLS, CA, and additionally SILWET L7607 (poly(ethylene oxide)-modified polydimethylsiloxane) was followed by polymerization of the styrene monomer to produce hybrid latexes. (36,37) Conditions of the emulsification (shear, stabilizers) and polymerization (initiator type: oilsoluble vs water-soluble) were varied. Extremely broad particle size distributions (20 nm to 2 [micro]m) were produced when the Manton-Gaulin homogenizer was employed as the primary means of creating the initial droplets with many particles being produced outside the miniemulsion size range (50-500 nm). Narrower size distributions were achieved employing a rotor-stator type homogenizer.
The breadth of the initial droplet size distribution was found to strongly affect the composition of the resulting hybrid particles after polymerization. A density gradient column (varying sucrose content) was used to separate particles of differing compositions (densities) whereby populations of small particles (D < 100 nm) comprised mostly of polystyrene and large particles (D > 250 nm) comprised largely of the Kraton rubber, were found. The redistribution of the styrene monomer during the polymerization was explained by the faster consumption of monomer in the smaller particles (formed by particle nucleation outside the droplets) drawing the styrene from the larger Kraton-containing species, thereby concentrating the rubber in them. More homogeneous compositions were found when an oil-soluble initiator (2,2'-azobis-2-methylbutyronitrile, AMBN) was used in place of the water soluble potassium persulfate, and this generally increased with higher levels of the initiator. Other effects included reduced induction periods, molecular weights, coagulum, and gel contents.
The mechanical properties of compression-molded 20 wt% Kraton containing hybrid latexes had elongations of 30% or more, indicating that this miniemulsification/polymerization method holds promise as a means of preparing rubber-toughened plastics.
Encapsulation--As envisioned in the initial discussions regarding the application of the miniemulsion technique (Figure 1), encapsulation of pigment was proposed, brought about by dispersion of a pigment in monomer, followed by miniemulsification and polymerization. This was only realized in recent years where Erdem et al. dispersed Ti[O.sub.2] particles in styrene monomer containing a stabilizer for the Ti[O.sub.2] (OLOA 370, polybutene-succinimide pentamine), plus HD costabilizer, and 1% polystyrene, the latter being added with the intent of ensuring that all droplets were nucleated ("enhanced" nucleation). (38,39) Following polymerization, the resulting latex particles were separated in a sucrose density gradient column (DGC) and characterized. (40)
For a system containing 3 wt% Ti[O.sub.2] (hydrophobic; trimethoxy octyl silane treated; 21-29 nm) based on styrene monomer, encapsulation efficiencies of 88% of the Ti[O.sub.2] and 73% of the PS (percentages involved in the encapsulation) were achieved. These numbers indicate that about 17% of the Ti[O.sub.2] remained unencapsulated and about 27% of the polystyrene was divided among latex particles containing no Ti[O.sub.2]. In addition, the encapsulated particle size was found to increase with increasing density (found by separating the sucrose layers) and the shapes of the particles became irregular at the higher densities. The former results remain unexplained. This method was further extended to a film forming system (styrene/n-butyl acrylate copolymer) achieving only poorer encapsulation efficiencies (64% of the Ti[O.sub.2] and 39% of the copolymer). (41) These results, plus the low level of loading (3 wt% based on polystyrene) prompted a change in strategy, namely, emulsification of the Ti[O.sub.2] in a polymer solution.
To achieve higher and more realistic levels of Ti[O.sub.2] loading, miniemulsification with the intent of preparing encapsulated artificial latexes was studied. (41) Here, pigment volume concentrations (PVC: volume of pigment/volume of pigment plus polymer) were varied from 11-70% in preparing artificial latexes. In this method, styrene/n-butyl acrylate copolymer (prepared by emulsion polymerization) was dissolved in toluene along with HD costabilizer. The Ti[O.sub.2] (hydrophilic; 20% rutile; 80% anatase; 29 nm) and its stabilizer (Solsperse 32,000; polyamine-polyester comb polymer) were added to the mixture and sonified to produce the Ti[O.sub.2] dispersion. This dispersion was then emulsified in an aqueous SLS solution via sonification to produce the miniemulsion. Removal of the toluene solvent via vacuum stripping produced the final artificial latex.
[FIGURE 10 OMITTED]
As in the previous work, characterization of the encapsulation efficiency was important. However, the usual sucrose solutions could not be applied as these particles would have a much higher density (e.g., for 30% PVC, [rho] = 2.97 g/cc) than those using only 3 wt% Ti[O.sub.2] (0.73% PVC). Instead, sodium polytungstate solutions were made varying in density up to 3.1 g/cc. Density gradient columns were again used to effect separation of the particles. In this work, no particles were found in the highest density layer, indicating that no Ti[O.sub.2] went unencapsulated (note: [[rho].sub.[TiO.sub.2]] = 4.1 g/cc), an improvement over the miniemulsion polymerization method. Some polymer particles (1-2 wt% of the copolymer) were produced with no Ti[O.sub.2] inside them, although this is much smaller than in the preceding encapsulation by polymerization studies. Interestingly, in the DGC separations, often one or more of the density layers contained no observable quantity of particles, even though adjacent levels did. In the previous work with the sucrose DGC, all levels contained some amount of particles. This difference also remains unexplained. It is not known how the emulsification process could produce such discontinuities. One consistent result was that larger particles were found in the higher density segments of the DGC.
The artificial latexes produced in these studies were further characterized in terms of their ability to form films and create hiding power as is one of the goals of coatings formulations. Indeed, these poly(styrene-co-n-butyl acrylate) encapsulated Ti[O.sub.2] particles formed continuous films and achieved hiding power at film thicknesses that decreased with increasing loading (PVC). Mixtures of separate Ti[O.sub.2] and copolymer dispersions made by miniemulsification were unable to achieve the same results for the same compositions showing the advantage of the miniemulsification approach.
[FIGURE 11 OMITTED]
MINIEMULSION POLYMERIZATION AS AN ALTERNATIVE TO EMULSION POLYMERIZATION
This paper has presented, through examples, some of the unique abilities and applications of miniemulsions. Some of these could simply not be developed by ordinary emulsion polymerization processes while others in fact could be.
Compared to conventional emulsion polymerization, miniemulsion polymerization might prove advantageous when an additional degree of control is desired. Since particle formation takes place in the monomer droplets, this means that a certain amount of control can be exerted on the final particle size by exercising control over the droplet size. Although 1-to-1 copying is not operative except under rare conditions, this still allows a more direct control not available in the conventional process.
The presence of the costabilizer can affect the composition of the monomers in the various loci in copolymerizations and might be used to some advantage in regulating this composition in the particles. (42) This affects not only composition but also the microstructure of the particles. The greater swelling of the particles containing the costabilizer can affect both of these to significant degrees and can differ much from conventional processes.
High solids latexes (~60%) can be prepared utilizing miniemulsion polymerization. (43) The broad droplet size distribution is seen as an advantage here.
Overall, we believe that miniemulsions are on the verge of a wider technological exploitation. Despite some gaps in our fundamental understanding, much has been accomplished in 30 years of research, which may still be only at the beginning. Only time will tell.
Table 1--Swelling Capacity ([V.sub.f]/[V.sub.i]) of Droplets/Particles with Styrene Monomer as Function of [j.sub.2.] (14) [j.sub.2] [V.sub.f]/[V.sub.i] 1 4000 2 1350 5 355 10 125 [infinity] 4.5
We wish to acknowledge all of the unnamed students, post-docs, visiting scientists, professors, and colleagues who were instrumental and vital to our research in miniemulsions over the past 30 years. Without financial support from the National Science Foundation, the Emulsion Polymers Industrial Liaison Program, and Lehigh University, among others, none of the preceding work could have been accomplished.
(1) Ugelstad, J., El-Aasser, M.S., and Vanderhoff, J.W., "Emulsion Polymerization: Initiation in Monomer Droplets," J. Polym. Sci., Polym. Letters Ed., 11, 503 (1973).
(2) Chou, Y.J., El-Aasser, M.S., and Vanderhoff, J.W., "Mechanism of Emulsification of Styrene Using Hexadecyltrimethyl-Ammonium Bromide Cetyl Alcohol Mixtures," in Polymer Colloids II, Fitch, R.M. (Ed.), p. 599, Plenum, New York, 1980.
(3) Grimm, W.L., Min, T.I., El-Aasser, M.S., and Vanderhoff, J.W., "The Role of Low Concentrations of Ionic Emulsifier Fatty Alcohol Mixtures in the Emulsification of Styrene," J. Colloid and Interface Sci., 94 (2), 531 (1983).
(4) Lack C.D., "Emulsion Formation and Stabilization with Mixed Emulsifier Liquid Crystals," Ph.D. Dissertation, Lehigh University, PA, 1985.
(5) Goetz, R.J., "Formation and Stabilization of Oil-in-Water Miniemulsions," Ph.D. Dissertation, Lehigh University, PA, 1990.
(6) Hessel, J.F., "Phase Behavior and Properties of Lamellar Surfactants," Ph.D. Dissertation, Lehigh University, PA, 1993.
(7) Asua, J.M., "Miniemulsion Polymerization," Prog. Polym. Sci., 27, 1283 (2002).
(8) Antonietti, M. and Landfester, K., "Polyreactions in Miniemulsions," Prog. Polym. Sci., 27, 689 (2002).
(9) Capek, I. and Chern, C.S., "Radical Polymerization in Direct Mini-Emulsion Systems," Adv. Polym. Sci., 155, 101 (2001).
(10) Sudol, E.D. and El-Aasser, M.S., "Miniemulsion Polymerization," in Emulsion Polymerization and Emulsion Polymers, Lovell, P.A. and El-Aasser, M.S. (Eds.), Chapter 20, p. 700, J. Wiley & Sons Ltd., England, 1997.
(11) El-Aasser, M.S. and Miller, C.M., "Preparation of Latexes Using Miniemulsions," in Polymeric Dispersions: Principles and Applications, Asua, J.M. (Ed.), p. 189, Kluwer Academic Publishers, 1997.
(12) Anderson, C.D., Sudol, E.D., and El-Aasser, M.S., "50 nm Polystyrene Particles via Miniemulsion Polymerization," Macromolecules, 35, 574 (2002).
(13) Morton, M., Kaizerman, S., and Altier, M.W., "Swelling of Latex Particles," J. Coll. Sci., 9, 300 (1954).
(14) Ugelstad, J., "Swelling Capacity of Aqueous Dispersions of Oligomer and Polymer Substances and Mixtures Thereof," Makromol. Chem., 179, 815 (1978).
(15) Anderson, C.D., "Nanosize Latex Particles via Miniemulsion Polymerization," Ph.D. Dissertation, Lehigh University, PA, 2002.
(16) Erdem, B., Sully, Y., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Determination of Miniemulsion Droplet Size via Soap Titration," Langmuir, 16, 4890 (2000).
(17) Hansen, F.K. and Ugelstad, J., "Particle Nucleation in Emulsion Polymerization. IV. Nucleation in Monomer Droplets," J. Polym. Sci., Polym. Chem., 17, 3069 (1979).
(18) Choi, Y.T., El-Aasser, M.S., Sudol, E.D., and Vanderhoff, J.W., "Polymerization of Styrene Miniemulsions," J. Polym. Sci., Polym. Chem. Ed., 23, 2973 (1985).
(19) Delgado, J., El-Aasser, M.S., and Vanderhoff, J.W., "Miniemulsion Copolymerization of Vinyl Acetate and Butyl Acrylate. I. Differences Between the Miniemulsion Copolymerization and the Emulsion Copolymerization Processes," J. Polym. Sci., Part A: Polym. Chem., 24, 861 (1986).
(20) Li, M., "Preparation of Urethane/Acrylic Hybrid Miniemulsion Nanoparticles," Ph.D. Dissertation, Lehigh University, PA, 2002.
(21) Chern, C.S. and Chen, T.J., "Miniemulsion Polymerization of Styrene Using Alkyl Methacrylates as Reactive Cosurfactants," Coll. Poly. Sci., 275, 546 (1997).
(22) Bailleul-Lefebvre, V., Sudol, E.D., and El-Aasser, M.S., unpublished data.
(23) Miller, C.M., Venkatesan, J., Silebi, C.A., Sudol, E.D., and El-Aasser, M.S., "Characterization of Miniemulsion Droplet Size and Stability Using Capillary Hydrodynamic Fractionation," J. Colloid and Interf. Sci., 162, 11 (1994).
(24) Pan, G., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Nitroxide-Mediated Living Free Radical Miniemulsion Polymerization of Styrene," Macromolecules, 34 (3), 481 (2001).
(25) Miller, C.M., Sudol, E.D., Silebi, C.A., and El-Aasser, M.S., "Miniemulsion Polymerization of Styrene: Evolution of Particle Size Distribution," J. Polym. Sci., Part A: Polym. Chem., 33, 1391 (1995).
(26) Miller, C.M., Sudol, E.D., Silebi, C.A., and El-Aasser, M.S., "Polymerization of Miniemulsions Prepared from Polystyrene in Styrene Solutions. 2. Kinetics and Mechanism," Macromolecules, 28 (8), 2765 (1995).
(27) Miller, C.M., Blythe, P.J., Sudol, E.D., Silebi, C.A., and El-Aasser, M.S., "Effect of the Presence of Polymer in Miniemulsion Droplets on the Kinetics of Polymerization," J. Polym. Sci., Part A: Polym. Chem., 32, 2365 (1994).
(28) Miller, C.M., Sudol, E.D., Silebi, C.A., and El-Aasser, M.S., "Polymerization of Miniemulsions Prepared from Polystyrene in Styrene Solutions. 3. Potential Differences Between Droplets and Particles," Macromolecules, 28 (8), 2772 (1995).
(29) Blythe, P.J., Klein, A., Sudol, E.D., and El-Aasser, M.S., "Enhanced Droplet Nucleation in Styrene Miniemulsion Polymerization. 1. Effect of Polymer Type in Sodium Lauryl Sulfate/Cetyl Alcohol Miniemulsions," Macromolecules, 32 (21), 6944 (1999).
(30) Blythe, P.J., Klein, A., Sudol, E.D., and El-Aasser, M.S., "Enhanced Droplet Nucleation in Styrene Miniemulsion Polymerization. 2. Polymerization Kinetics of Homogenized Emulsions Containing Predissolved Polystyrene," Macromolecules, 32 (21), 6952 (1999).
(31) Blythe, P.J., Klein, A., Sudol, E.D., and El-Aasser, M.S., "Enhanced Droplet Nucleation in Styrene Miniemulsion Polymerization. 3. Effect of Shear in Miniemulsions that Use Cetyl Alcohol as the Cosurfactant," Macromolecules, 32 (13), 4225 (1999).
(32) Prodpran, T., Dimonie, V.L., Sudol, E.D., and El-Aasser, M.S., "Nitroxide-Mediated Living Free Radical Miniemulsion Polymerization of Styrene," Macromol. Symp., 155, 1 (2000).
(33) Pan, G., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Living Free Radical Miniemulsion Polymerization of Styrene," in Polymer Colloids: Science and Technology of Latex Systems, Daniels, E.S., Sudol, E.D., and El-Aasser, M.S. (Eds.), ACS Symposium Series 801, American Chemical Society, p. 139, Oxford University Press, 2001.
(34) Pan, G., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Surfactant Concentration Effects on Nitroxide-Mediated Living Free Radical Miniemulsion Polymerization of Styrene," Macromolecules, 35 (18), 6915 (2002).
(35) Pan, G., "Living/Controlled Free Radical Miniemulsion Polymerization of Styrene," Ph.D. Dissertation, Lehigh University, PA, 2002.
(36) Jeong, P., "Hybrid Composite Latexes," Ph.D. Dissertation, Lehigh University, PA, 2000.
(37) Jeong, P., Dimonie, V.L., Daniels, E.S., and El-Aasser, M.S., "Hybrid Composite Latexes," in Polymer Colloids: Science and Technology of Latex Systems, Daniels, E.S., Sudol, E.D., and El-Aasser, M.S. (Eds.), ACS Symposium Series 801, American Chemical Society, p. 357, Oxford University Press, 2001.
(38) Erdem, B., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Encapsulation of Inorganic Particles via Miniemulsion Polymerization I. Dispersion of Titanium Dioxide Particles in Organic Media Using OLOA 370 as Stabilizer," J. Polym. Sci., Part A: Polym. Chem., 38 (24), 4419 (2000).
(39) Erdem, B., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Encapsulation of Inorganic Particles via Miniemulsion Polymerization II. Preparation and Characterization of Styrene Miniemulsion Droplets Containing Ti[O.sub.2] Particles," J. Polym. Sci., Part A: Polym. Chem., 38 (24), 4431 (2000).
(40) Erdem, B., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Encapsulation of Inorganic Particles via Miniemulsion Polymerization III. Characterization of Encapsulation," J. Polym. Sci., Part A: Polym. Chem., 38 (24), 4441 (2000).
(41) Al-Ghamdi, G.H., "Encapsulation of Inorganic Particles via Miniemulsification and Film Formation of Resulting Composite Latex Particles," Ph.D. Dissertation, Lehigh University, PA, 2003.
(42) Kitzmiller, E.L., Miller, C.M., Sudol, E.D., and El-Aasser, M.S., "Miniemulsion Polymerization: An Approach to Control Copolymer Composition," Macromol. Symp., 92, 157 (1995).
(43) Leiza, J.R., Sudol, E.D. and El-Aasser, M.S., "Preparation of High Solids Content Poly(n-butyl acrylate) Latexes Through Miniemulsion Polymerization," J. Appl. Polym. Sci., 64, 1797 (1997).
Mohamed S. El-Aasser and E. David Sudol--Lehigh University*
*Emulsion Polymers Institute, Bethlehem, PA 18015.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||2002 Roy W. Tess Award in Coatings|
|Author:||Sudol, E. David|
|Date:||Jan 1, 2004|
|Previous Article:||2002 Roy W. Tess Award.|
|Next Article:||Optimizing nitroxide-mediated miniemulsion polymerization processes.|